Emulsion polymerization of butadiene : a kinetic study · EMULSION POLYMERIZATION OF BUTADIENE A...
Transcript of Emulsion polymerization of butadiene : a kinetic study · EMULSION POLYMERIZATION OF BUTADIENE A...
Emulsion polymerization of butadiene : a kinetic study
Weerts, P.A.
DOI:10.6100/IR332572
Published: 01/01/1990
Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
Citation for published version (APA):Weerts, P. A. (1990). Emulsion polymerization of butadiene : a kinetic study Eindhoven: Technische UniversiteitEindhoven DOI: 10.6100/IR332572
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?
Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Download date: 23. Jun. 2018
EMULSION POLYMERIZATION
OF BUTADIENE
A KINETIC STUDY
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de
Teehnisc;he Universiteit Eindhoven, op gezag
van de Rec;tor Magnificus, prof. ir. M. Tels, voor
een commissie aangewe:<:en door het College
van Dekanen in het openbaar te verdedigen op
vrijdag 8 juni 1990 te 16.00 uur
door
PETRUS ANTONIUS ~ERTS
geboren te Breda
dl',IH w.III·I~ 11'-"",,'1 "I.,,·, .. ·\ •• ~\·, '. 1",I""".d
Dit proefschrift i~ goedgckcurd door
de promotoren: prof. dr. ir. A.L German
prof. dr. R.G. Gilbert
en de copromotor: dr. W.A.B. Donners
The author is indebted to DSM Research, Geleen, The Netherlands, for
supporting this work.
Aan Marian, Renee en Pim
Chapter 1
1.1 Short historic overview
1.2 Aim of this investigation
1.3 Survey of thesis
References
Chapter 2
2.1 Emulsion polymerization
Contents
Introduction
Emulsion Polymerization
2.2 Emulsion polymerization kinetics
2.2.1 Average number of radicals per particle
2.2.2 Particle number
References
Chapter 3 Experimental Procedures and Techniques
3.1 Standard polymerization recipes
3.2 Reactor design and experimental procedures
33 Emulsifier
3A Initiator
3.5 Buffer
Contents
1
3
4
6
7
8
9
13
14
15
17
19
20
23
Contents
3.6 Chain transfer agent and gel analysis
3.7 Particle size analysis
3.8 Polybutadiene
References
Chapter 4 Polymerizations with Rosin Acid Soap
4_1 Introduction
4.2 Effect of the initiator concentration
4.3 Effect of the emulsifier concentration
4.4 Conelusion~
References
Chapter 5 Polymerizations with Sodium Dodecylsulfate
5.1 Introduction
5.2 Effect of the emulsifier concentration
5.3 Effect of the sodium ion concentration
5.4 Effect of the initiator concentration
5.5 Comparison between dresinate 214 and sodium dodecylsulfate
5.6 Conel usions
References
Chapter 6
6_1 Introduction
62 Agitation
6.3 Monomer to water ratio
6.4 Conclusions
References
Physico-Chemical A~pects
24
26
2S
31
33 34
37 45
46
47
48
51
57
61
65
66
67
68
80
86
87
Contents
Chapter 7 Kinetics of the Emulsion Polymerization of Butadiene
7.1 Introduction 89
7.2 Interval II analysis 90 7.3 Interval In analysis 94
7.4 Effect of the initiator concentration 97
7.5 Radical loss mechanisms 99 7.6 Conclusions 104
Appendix. Calculation of the saturation monomer concentration 105 References 107
Chapter 8 Effects of Thlol5
8.1 Introduction 109
8.2 Thiol as chain transfer agent 110
8.3 Promoting effect of thiols llZ
8.4 Emulsifier-free polymerizations 114
8.5 Radical desorption 118
8.6 Conc1usion$ 119
Appendix. Anomalous effects with dresinatc 214 as emulsifier 1Z0
References 123
Summary 125
Samenvatting 127
Glossary of Symbols 129
Curriculum Vitae 133
Oank:woord 134
Introdnf;'ion l
Chapter 1 Introduction
1.1 Short historic overview
The first attempts to develop an emulSion-like process were undertaken in
Gennany prior to and during World War I, out of necessity to cope with the
shortage of natural rubber. Nowadays the processes disclosed in several patents
granted to Farbenfabriken Bayer A.G.1,<J,4) would be referred to a.~ suspension
polymerizations; diene-monomers were polymerized in aqueous dispersions with
egg albumin, starch or gelatin as stabilizers.
The first reference to a process which can be rCg'<lorded as a true emulsion
polymerization appeared in 1927, a patent granted to DinsmoreS) working for the
Goodyear Tire & Rubber Company. It describes the thermal polymerization over
a period of six months of aqueous emulsions of various diene.monomers, with
oleate salts and egg albumin a.~ stabilizers. Luther and Heuck6) of LG_
Farbenindustrie A.G. in 1932 were the first to introduce initiators to facilitate
rapid polymerization. Although these first emulsion polymerizations were
certainly not commercially viable, they led to the development of a complete
industry over the nellt two decades.
Under the impulses of World War II the Synthetic Rubber Program in the
United States, under Goordination of the 'Office of Rubber Re~erve' (ORR), led
to the successful production of butadiene(75)-styrene(25) copolymer (socalled
GR-S nIbber) as a general purpose rubber. At the end of 1945 the total OR-S
production had already reached a level of 7.5 HP tons;7) natural rubber would
never regain its leading position.
2 Chapter 1
Under the synthetic rubber program research was more of an applied than
a theoretical nature, and focused on the development of copolymers of
butadiene with styrene or acrylonitrile, as these had qualities superior to that of
the polyhutaJiene homopolymer. Yet, much of the fu))damental knowledge about
the emulsion homopolymerization of butadiene stems from this early work.
Excellent reviews8) arc available, although some of the work was only
declassified in 1980. However, little attention has been paid to butadiene since
those early days of re~earch, eventhough the importance of polybutadiene
containing polymers and resins has increased enormously (e.g. high impact
materials, coatings, adhesives). A brief summary of relevant literature after 1950
will he given below.
In the early fifties Morlon et al. published a series of anicles about the
cross-linking behaviour of polybutadiene,?) and the use of several initiator
systems in the emulsion polymerization of hutadiene. IO,I1,12) The Only values
reported for the propagation rate coefficient (kp) stem from this work,1I) but at
best represent an approximation of the true values since they were determined
with ab initio polymerizations at low temperatures (0 - 30"C), involving poly
disperse latexes sized by the method of soap titration.
In 1964 a Ph. D, Thesis by BhakuniB ) appeared on the "Kinetics of the
Persulfate-Mercaptao Emulsion Polymerization of Butadiene". Several notcn
worthy conclusions were real;hed: the nature of the emulsifier influences real;tion
kinetics, network formation within the latex particles has no influence On the
saturation monomer solubility in the particles, while the possibility of radical
desorption was mentioned.
More recently Wendler et aI. 14,15,16,17) reported about the emulsion
polymerization of butadiene in the presence of excessive amounts of
bis(alkyloxythiocarbonyl)disulfanes in order to prepare low molecular weight
oligobutadienes. Rt:sults were compared with the Smith-Ewart tht:ory, indicating
a strong deviation from the 'ideal' case II behaviour. The average number of
radicalS pCI' particle (n) was supposed to be larger than 0.5.
Of some interest is the radiation-induced emulsion polymerization of
butadiene. IS) Polymerizations were found to be much slower than anticipated
Introduction 3
from the teported kp value/I) but comparable with the corresponding persulfate
initiated polymerizations. A profound effect of the mode of agitation wa~ found,
indicating a diffusion-controlled process.
Results from Russian wotkers have been reviewed by Eliseeva et a1.. 19)
When increasing the initial butadiene/5tyrene ratio from zero to one in the
emulsion copolymerization of butadiene and styrene, the final particle number
increases, whereas the overall polymerization rate decreases. It is stated that this
effect can only be partially explained by the difference in 1)" so that other yet
undetermined factors must be operative. Similar results were found lJy others.lJ)
Altogether, literature on butadiene emulsion polymerization is sparse_ The
chronic lack of reliable rate coefficients makes a thorough interpretation of
kinetic data difficult. Despite the long history of emulsion (co)polymerization
of butadiene, the process definitely merits further investigation.
1.2 Aim (If this investigation
The investigation described in this thesis aimS at a better understanding of the
kinetics of the ab initio emulsion polymerization of butadiene. Any emulsion
polymerization system is complex due to its heterogeneous and colloidal nature.
In many studies attention is focused on the Chemical parameter~ that influence
the process (usually the initiator and emulsifier concentration). In this study a
complete survey of all relevant parameters was intended. The following
parameters were considered, classified according to their nature:
Chemical:
Colloidal:
type and concentration of initiator
amount of chain transfer agent
monomer to water ratio
intensity of agitation
type and concentration of emulsifier
overall cation concentration
4 Chapter I
As a starting point, a typical industrial recipe was chosen, using the
frequently empl(lyed emulsifier dresinate 214. In order to establish whether the
anomalou~ hchaviollf found was related to specific experimental conditions,
polymerizations with the well-defined emulsifier sodium dodecyhulfate were
performed under comparable conditions. No major recipe-related phenomena
were found, so a generalized semi"quantitative description could be given, which
in turn led to novel insights and a reinterpretation of results reported by others.
l.3 Survey of thesis
Major theoretical aspects of the mechanisms and kinetics of emulsion
polymerization relevant to this inveStigation are briefly discussed in Chapter 2.
Chapter 3 deals with the experimental procedures and techniques used
throughout this investigation, and typical aspects of various recipe ingredients
(e.g. composition of emulsifier, initiator de~omposition and gel analysis).
rn Chapter 4 the ab initio emulsion polymerization of butadiene with the
industrial grade emulsifier dresinate 214 was investigated. Attention is focused
on the kinetic effects of the initiator and emulsifier concentration. The main
body of this chapter has already been publisbed.20)
In Chapter 5 the well-defined research grade emulsifier sodium dodecyl
sulfate was used in studying the effects of initiator and emulsifier concentration.
Since the performance of this emulSifier is virtually pH-independent, the ionic
strength was varied by changing the amount of sodium carbonate in the recipe.
A profound effect on colloidal stability was found; under the experimental
conditions ~hoseIl particle nucleation is coagulative in nature.
As industrial batCh processes are often performed at the highest monomer
to water ratio feasible, the effect of this parameter was studied in Chapter 6.
Furthermore, the effect of agitation was considered with regard to the
emulsification of the monomer, and its impact on reaction kinetics.
Introduction 5
The ab initio experiments described in the preceding chapters contain
valuable information about the kinetics of butadiene in emulsion polymerization.
A kineti, analysis of interval II and interval ru is presented in Chapter 7,
showing conclusively that the present system should be treated as a zero-one
system, with n « OS
In Chapter 8 the role of thiols with a low water solubility is evalu:3.ted.
Primarily they are added to control the extent of branching and cross-linking,
and thus the amount of insolUble gel formed. In this respect they behave as ie.kal
chain transfer agents, as no gross effects on polymerization kinetics were found
(except maybe at a high dresinate 214 to butadiene ratio). The socalled
promoting effect of thiols has also been ,onsidered. A minor part of this chapter
has been published in Polymer Communications.21) The chapters 5,6 and 8,22) as
well as chapter 7'1:» will be submitted for publication.
Parts of this work have been presented at the International Symposium on
Free Radical Polymerization (Santa Margherita Ligure, May 1987), the 2nd
International Symposium on Copolymerization and Copolymers prepared in
disperse media (Lyon, April 1989), the Rolduc Polymer Meeting4 (K.erkrade,
April 1989) and the Polymer Latel! HI Conference (London, June 1989)_
6 Refenmces
Ger. 250,690 (1909), Farbenfabriken Bayer A.G., Inv.: F. Hofmann, K. Dclbriick
2 Ger. 254,672 (1912), Farbcnfabriken Bayer A.G., Inv.: F. Hofmann, K. Delbriick
3 Ger. 255,129 (1912), Farbenfabrikcn Bayer A.G., lnv.: F. Hofmann, K. Delbruck
4 US 1,149,577 (1915), Farbenfabriken Vorm. Friedr, Bayer & Co., Inv.: K. Gottlob
5 US 1,732,795 (1929), The Goodyear Tire & Rubber Co., Inv.: R.P. Dinsmore 6 US 1,860,681 (1932), I.G. Farbenindustrie A.G., lnv.: M. Luther, C Heuck 7 R.F. Dunbrook, in "Synthetic Rubber", edited by G,S, Whitby, C.C. Davi~,
R.F. Dunbrook, John Wiley & Sons, New York 1954, p. 51 8 F.A. Bovey, I.M. Kolthoff, AJ. Medalia, E.J. Meehan, ''Emulsion
Polymerization", Interscience Publishers, New York 1955 9 M. Morton, P.P. SaIatiello,l Polym Sci. 6, 225 (1951)
10 M. Morton, P.P. Salaliello, H. Landfie!dJ f'olym. Sci 8, 111 (1952) 11 M. Morton, P.P. SalatieIlo, H. Landfield,l Polyrn. Sci. 8, 215 (1952) 12 M. Morton, P.P. Salatiello, H. Landfield, Ind. Eng. Chern. 44, 739 (1952) 13 R.s. Bhakuni, Ph. D. Thesis, University of Akron (1964); Chern Abstr. 62,
13351b (1965) 14 K. Wendler, N. Karim, M. Fedtke, Plaste Kautsch. 30, 247 (1983) 15 K. Wendler, L Pielert, M. f.'edtke, Plaste Kautsch 30, 438 (1983) 16 K. Wendler, R. Wiirtenberg, M, fedtke, Plasle Kautsch. 31, 367 (1984) 17 K. Wendler, H. Elsner, W.D. Hergelh, M. Fedtke, Plasre Kautsch. 32, 128
(1985) 18 K. [shigure, T. O'Neill, E.P. Stahe\, V. Stannett, J. MacromoL Sci. - Chern"
AS, 353 (1974) 19 V.I. Eliseeva, S.S. lvanchev, S.l. Kuchanov, A.V, Lebedev, "EmuLrion
polymerization and its applications in industry", Khimiya, Moscow 1976; Engli~h trans!.: SJ. Teague, Consultants Bureau, New York 1981
20 P"A Weerts, J.LM" van der L)os, AL German, Makromol. Chern. )90, 777 (1989)
2J P.A Weerts, J.L.M. van der Loos, AL. German, Polym. Comm. 29, 278 (1988)
22 P.A. Weerts, J,L.M. van der wos, AL. German, Makromol. Chern., submitted
23 P.A Weerts, AL. German, R.G. Gilbert, Macromolecules, submitted
Emulsion PolymerizatioB 7
Chapter 2 Emulsion Polymerization
2.1 Emulsion polymerization
In contrast to bulk and solution polymerization, emulsion polymerizations
are heterogeneous reaction systems, initially consisting of a continuous aqueous
phase and finely dispersed monomer droplets. Systems with a discontinuous
aqueous phase also occur, referred to as inverse emulsion polymerizations. The
product, a colloidal dispersion of submicron polymer particles, is called a latex.
Water-soluble radical initiators are mostly used in emulsion polymerizations,
and reaction takes place mainly in the monomer-swollen latex particles; this in
contrast to the related process of suspension polymerization where oil-soluble
initiators are employed to initiate polymerization in the monomer droplets.
Emulsifiers are often added to assist in the particle formation process, and to
enhance colloidal stability of the latex particles. Optionally other ingredients may
be included in the recipe, such as chain transfer agents to reduce the molecular
weight of the polymer, or electrolytes to influence the colloidal and rheological
properties of the latex.
The advantages of emulsion polymerization are well-known. (1) The heat
generated during the exothermic polymerization is quite easily dissipated via the
aqueous phase. (2) The rate of polymerization is greater than in the
corresponding bulk process, (3) while at the same time the molecular weight of
the polymer formed is considerably higher. (4) The latter can readily be
controlled by addition of chain transfer agents. The resulting latex is wattr-
8 Cbapter 2
based, which is advantageous in certain applications (e.g. paints), but it also
contains emulsifier and initiator residues which are hard to remove and may
affect the polymer properties adversely.
It is convenient to divide the course of an emulsion polymerization into
three distinct intervals. Interval I is the initial stage where particle formation
takes place. Several mechanism~ of particle nucleation have been proposed,!)
involving the emulsifier micelles,2) the aqueous phase,J) or the monomer
dtoplets;4) some mechanisms take account of the colloidal instability of the latex
particles.M,'T) Interval n is characterized by a constancy of particle number, while
polymerization in the particles proceeds in the presence of a separate mOnomer
pha.~c. The beginning of interval II is sometimes taken as the conversion where
the emulsifier concentration drops below its critical micelle concentration.
Interval III begins with the disappearance of monomer droplets, after which the
monomer concentration in the particles starts to decrea.~e continuously.
2.2 Emulsion polymerization kinetics
Emulsion polymerization is a particular case of free radical polymerization.
The kinetics and mechanisms involved are highly complicated, since events
occurring in several phases, and exchange of radicals between the phases need
to be taken into account. Models and theories dealing with these aspects of
emulsion polymerization are numerous, but excellent reviews are avaHable:·~ 1,8,9)
A brief discussion of emulsion polymerization kinetics in order to introduce a
theoretical framework in the context of this investigation is thus considered
sufficient. Details can be found in the references cited.
In general, the rate of free radical polymerization Rp<>I is given by;
RF<>I = ~ [M) [RO] (2.1)
Emulsion Polymemation 1)
where k., is the propagation rate coefficient, [M] and [R-] the monomer and
free radical concentrations, respectively. This equation is not convenient to use
because of the unknown radical concentration [R-]. In emulsion polymeri~tion
the overall polymerization rate can be taken as a summation of those in e~h
individual latex particle, leading to:
(2.2)
where c,.. is the monomer concentration within the latex particles, n the average
number of radicals per particle, N the particle number per unit volume of
aqueous phase, and NA¥ Avogadro's number_ The parameters most difficult to
predict are n and N.
2.2.1 Average number of radicals per particle. A set of population balance
equations can be given for the the number of particles Nn containing n radicals:
+ Ie. [(n + 2)(n + 1)Nn ' 2 - n(n - l)N.J (2_3) v
where p. is the rate of entry of free radicals into the particles (absorption), 1.")
the rate coefficient for exit of radicals from the particles ("desorption"). k1 the
rate coefficient for bimolecular termination within the particles, and v the
volume of a monomer-swollen particle, all in appropriate units. The rate of
radical production in the aqueous phase Pi' which was sometimes used in earlier
equation~ Cor N, is equal to p. only when radical exit is negligible and the
initiator efficienCy 100 % .
• ) Smith and EwartW) applied k"A/v instead of k, where A is the surface
of a monomer-swollen particle, and k" an elCit rate coefficient.
lO Chapter 2
The first term5 on the right hand sIde of equation (2.3) account for cbanges
in No due to radical entry, the second terms for radical exit, and the third ones
for bimolecular termination.
Smith and EwareO) were the first to obtain some limited solutions for the
~ystem represented by equation (2.3), based on a steady-state analysis (i.e. taking
dNJdt = 0) for monodisperse latexes. Three special cases were distinghLlished.
Case 1, with it " l. This situation occurs if the exit rate is great as
compared with the (:ntry rate, i.e. (PaiN) « k. Under these conditions only the
first of the recurrence relationships of equation (2.3) needs to be considered, as
particles containing two or more radicals may he neglected:
Since No '" N, it follows that Nl "" (P.lk)
polymerization RI'<" can be written as:
o
n.N; thus the overaH rate of
(2.4)
Case 2, with ii = 0.5. In this case, which is most generally known as the
Smith-Ewart theory, the following conditions need to be satisfied simultaneously:
(1) radical exit is negligihle, (2) and bimolecular termination is instantaneous
when a second radical enters a particle, i.e. k « (p./N) « (~/v). It follows
trivially that No "' N1, so that n '" Nd(No + N1) "' 0.5. Only in this case is Rpoj
pwpoTtional to the particle number N:
(2.5)
Cast 3, with ii »1. This situation will occur when the entry rate is much
greater than the rate of bimolecular termination, i.e. (ktfv) « (p./N). Smith and
Ewa.rt neglected radical exit in their treatment of this case. With a sufficiently
Emulsion Polymerization 11
large ii, the steady-state condition is p.lN ::: 2~ i'i Z Iv. Since the totaL volume of
polymer per unit volume of aqueous phase V ::: N.v, the rate ~I becomes;
(2.6)
The Smith-Ewart case 1 and 2 situations are conveniently encompassed in
the socalled zero-one sYStem, where only zero or one free radical per particle
need to be considered. The main condition for such a system is that the rate of
bimolecular termination is considcrably greater than those for radical entry Or
exit. Two radicals cannot coexist in a particle because they terminate virtually
instantaneously, viz. n ::: 0.5. For convenience, the pseudo-fIrst-order entry rate
coefficient (J is introduced to repLace (J./N in equation (2-3), In the steady-state
the population balance is:
so that (2.7)
Smith-Ewart case I behaviour is characterized by k » p, thus Ii '" p/k, while in
case II k« p, inevitably leading to Ii .. 0.5.
StockmayerllJ was the first to give a general solution for equation (2.3)
involving Bessel functions, which was later modified by O'TooieI2) to give:
(2.8)
where I represents Bessel functions of the firM kind, and m and a dimensionless
parameters defined as:
m k.v/Is a (&)0.5
Ci p.v/Nt,
12 Chapter 2
The parameter m is a measure of the exit rate relative to the Tate of
bimolecular termination (e.g. In = 0, no exit), while a represents the rate of
radical entry V5, the Tate of bimolecular termination. Utilization of equation
(2.8) is not ~tl'aightforwaTCJ because Po cannot be obtained unambiguou~ly, as
it is influenced by the Tates of radical production in the aqueous phase,
termination in the aqueous phase, and reentry of exited free radicals. Ugelstad
et alY) derived a general method of calculating n, taking the abovementioned
aqueous phase events fully into account. The following equation was obtained:
where: c;;'
y
,,' + mn - Yci
Pj'v/Nk,
2Nk,k"'4/k• \.
(2.9)
with Pi the nlte (If Tadical production in the aqueous phase, k'oq the termination
rate coefficient in the aqueous phase, and k. the absorption rate coefficicnt.
10'
10' 10' ~. = p,v/Nk,
Figure 2.1. Ugelstad plot for Y '" 0.13)
EmulSion Polymerization 13
Numerical methods were used to obtain solutions for equations (2.8) and (2.9).
Typical results are shoWIJ in Figure 2.1, for the case of negligible aqueous phasE
tennination (Y = 0). The three special Smith-Ewart cases are fully encompassed
in this figure; the straight horizontal line with m = 0 represents case 2 (n =
0.5), the curves on the lower left side with m > 0 include case 1 (ii «0.5), while
the line in the upper right side gives ca:;e 3 (n » 0.5).
2.2.2 Particle number. The Smith-Ewart modelIO) for the predktion of the
particle number i~ based on their case 2 growth kinetics (Le. a constant volume
growth rate), assuming a micellar particle nucleation mechanism as proposed by
Harkins. I4,IS) Particle nucleation supposedly stops when the total surface of the
particles becomes sufficient to adsorb all the emulsifier added. The particle
number N predicted by this nucleation model is given by:
(2.10)
where K is a constant between 037 and 0.53, () the volume growth rate of a
particle, A.:. the specific area per unit amoum of emulsifier, and IE] the
emulsifier concentration, all in appropriate units. Usually, Pi can be taken
proportional to the initiator concentration [I], so that N '" [I]0.4.[Et6.
It was later shown by RoeUi) that this power law is also consistent with a
homogeneous nucleation mechanism, where particles are nucleated by precipita
tion of oligomeric radicals in the aqueous phase. In fact, the two exponents are
predicted by any model that takes cessation of complete surface coverage by
emulsi.fier as the main criterion of the end of the parti.cle nudeation period.
However, experimentally observed exponents cover a wide range ofvalues,t7)
while they do not always obey a simple power law when plotted over a large
concentration range_IS) Nowadays it is realized that limited coagulatiOn (or
coalescence) of the latex particles should be taken into account; models based
on this concept are extremely complex, but also show a good agreement with
experimental findings.6,7,!9)
14 References
F.K. Han~en, J. Ugelstad, "Ptlrlicie Fonnation Mech(lJtl$ms", in: "Emulsion Polymerization", edited by I. Piirma, Academic Press, New York 1982
2 W.D. Harkins,l. Am. Chern. Soc. 69, 1428 (1947) 3 W.J. Priest, J. Phys. Chern. 56, 1077 (1952) 4 J. Ugelstau, rvts. El-Aasser, J.W Vanderhoff, J. Po/ym. Sci., Polym. Lett.
Ed. 11, 503 (1973) 5 FK Hansen, J. Uge.\stad, 1 Polym. Sci., Polym. Chern. Ed. 16, 1953
( 1978) 6 PJ. Fe~ncy. OJI. Napper, R.G. Gilbert, Macromolecules 17,2520 (1984) 7 Z. Song, G.W. Poehkin, 1. MacromoL Sci.·Chem. AlS, 403, 1587 (1988) 11 J. Ugelstad, FK Hansen, Rubber Chern. Technol. 49,536 (1976) 9 RG. Gilbert, D.H. Napper, 1. Macromol. Sci - Rev. MacromoL Chern
Phys. C23, 127 (1983) 10 w.v. Smith, R.H. Ewart, J. Chem. Phys. 16,592 (1948) 11 W.H. Stock mayer, 1. Po/ym. Sci. 24, 314 (1957) 12 J.T. O'Toole, J. Appl. Polym. Sci. 9, 1291 (1965) 13 1. Ugelstad, P.e. M5rk, J.O. Aasen,]. Polym. Sci., Part A-J S, 2281 (1967) 14 W.D. Harkins, J. Am. Chern. Soc. 69, 1428 (1947) 15 W.O. Harkins, J. Po/ym. Sci. S, 217 (1950) 1Ii c.P. Roe, Ind. Eng. Chern. 60, 20 (1968) 17 R. M. Fi teh, Brit. Polym. J. 5, 467 (1973) 18 N. Sutterlin, H.-J. Kurth, G. Markert, Makromol. Chern. 177, 1549 (1976) 19 JR RichanJo, J.P. CongaHdis, R.G. Gilbert, 1. Appl. Polym. Sci. 37,2727
(1989)
Experimental Procedures and Teehniqutls
Chapter 3 Experimental Procedures and Techniques
SUMMARY: The redpe ingredients, experimental set up, and rea.ction conditions used throughout this investigation arc outlined in detail. Procedures for gel-analysis and particle sizing are discussed, and some kinetically relevant properties of polybutadiene (e.g. gel content, density, monomer solubility) were determined. The decomposition of persulfate in the presence of the emulsifier dresinale 214 is 1.5 ' 2 times faster than in water, presumably due to oxidation of emulsifier components. The necessity to include a buffering substance in the recipe is illustrated; in its absence pH drop. below 3 and polymerization is severely retarded.
3.1 Standard polymerization recipes
15
The two standard recipes used throughout this study are shown in Table 3.1,
and were chosen in analogy with recipes found in industrial practicc.'·~ 1,1,3) For
comparison, the GR-S Or mutual recipe (for a butadiene homopolymerization)
developed during the synthetic rubber program in World War n fOr production
of general purpose butadiene(75)-styrene(25} rubber is shown in the last column.
The simila6ty between recipe 1 and the mutual recipe is striking; the spin
off of the synthetic rubber program is still relevant to current commercial
processes for the production of polybutadiene containing polymers and resins.
Recipe 1 is a typical industrial recipe with regard to the choice of the
emulsifier dresinate 214, being a mixture of surface-active components of natural
16 Chapter 3
origin (see section 3.3), with a quality and performance that may vary with
source and time. The number of batches dresinate 214 used throughout this
investigation was limited, while each series of experiments was always done with
one batch over a short period of time.
In order to avoid these problems. the analogous reCipe 2 was introduced,
employing the well-defined research grade emulsifier sodium dodecylsulfate.
Differences betwEen both standard recipes are not crucial and based on practical
considerations; potassium dodecylsulfate was found insoluble in water at room
temperature, so that a.ll potassium salts from recipe '1 were substituted by the
subsequent sodium salts on a molar basis. The overall cati(m concentration
(including the emulsifier contribution) in w;ipe 2 is [Na 'J '" 0.3 mol.L-\ while
in the standard recipe 1 [K' J = 0.4 mol.L-I.
In the following sections the experimental procedures and relevant
characteristics of the recipe ingredients will be discussed.
Table 3J Standard polymerization recipes in parts by weight.
Ingredient Recipe 1 Recipe 2 Mutual recipe
water 230 230 180 butadiene 100 100 100 dresinate 214 7.6 5') sodium dodecylsulfate 7.6 K+/Na' carbonateb
) 4.4 2.0 K + INa' persulfateb
) 0_8 0.7 0.3 t-dodecanethiol 0.7 0.7 0.5
(initiator] 13 mmol.L-1 6 mmoLL'! temperat,!Te 62"C 50"C pH 10.5 - 10.8
oj Initially a commercial emulsifier of 'ORR quality' was used, with sodium palmitate, stea.ate and oleate as main components; later this was frequently substituted by dresinates (see section 3.3).
b) In recipe 1 and the mutual recipe potassium salts, and in recipe 2 sodium salts were used.
Experimental Procedures and Techniques 17
3.2 Reactor design and experimental procedures
The butadiene (DSM Hydrocarbons, GeJeen, Holland) was pu..-ified by
condensing the vapour from a 27-litre storage vessel into a cooled l>teel recipient
(Figure 3.1, page 19). The polymerizations were performed in a stainless steel
reactor (from K. K Juchheim, Bernkastel-Kues, FRG) fitted with four baffle
plates located at 90" intervals and a twelve-bladed turbine impeller (Figure 3.2).
The reactor was charged under N] wHh all ingredients except the butadiene.
The system was freed from inert gases by flushing the gas cap with gaseous
butadiene, followed by evacuation. This procedure was repeated twice. LaMly.
the correct amount of heshly distilled liquid butadiene was added from a
weighed steel vessel.
The polymeriUltion temperature was kept constant within O.l°C, During
polymerization samples (ca. 8 mL) were taken with a high-pressure-proof
sytinge, and the conversion determined from the total solid content of the
sample. The polymerization rate RpOI') was taken as the slope of the linear
portion of the conversion-time curve (I.e. interval II).
Reactor dimensions in mm
impelJer diameter 60 reactor diameter 134 baffle diameter 13.4 blade diameter 18 reactor height 175 initial liquid heightb) 162 final liquid heightb) 139
b) Calculated for a standard recipe without sampling.
Figure 3.2 Cross section of the polymerization reactor .
• ) The steady-state polymerization rate Rpol is expressed in grams (or moles) of polymer formed per unit volume of water per second.
18 Chaptcl" 3
The effective volume of the reactor shown in Figure 3.2 is 2.36 litre. After
charging the reactor according to one of the ~landard recipes, the initial degree
of filling i~ 92 % A~ polymerization proceeos, the reaction volume decreases
due to tbe large difference in density hetween butadiene and polybutadiene.
Without any sample:; heing taken, the final degree of filling becomes 78 %,
Under atm()~pheric circum.tances butadiene is a gaseous mOnOmer (boiling
poim - 4,4"C); thus in a polymerizing system the amount in the gas phase needs
to be considered. Assuming ideal gas behaviour as a fir~t order approximation,
the amount of butadiene vapour at an initial degree of filling of 92 %, is 0.6 %
of the total amount charged. This constitutes a small systematic eHor for
gravimetric conversion data, a:; samples are taken from the emulsion.
Nevertheless, its magnitude is comparable with statistical errors involved in the
sampling procedure, as long as the gas space is minimized and exces.ive
sampling avoided. The effect on Rp<>1 is negligible, as Rp<>1 represents the :;Iope
of the conversion-time curve.
3_3 EIIlUlsiOCf
TwO emulsifiers have been used routinely: dresinate 214 (Hercules bv, Den
Haag, Holland) and sodium dodecylsulfate (SOS, purity > 99 %, Merck,
Darmstadt, PRG). Other emulsifiers employed o~casionally are potassium
stearat~ and oleate, formed in silu from stearic acid (purity > 98 %, Merck,
Schuchardt, f'RG) and oleic acid (Merck, Darmstadt, FRG).
Dresinate 2144;;) was us~d as a 15 weight % dispersion of potassium salts of
disproportiOtlated rosin acid soap in water. Rosin is a resinous substance
obtained from pine trees by solvent extraction of wood stumps, or collecting the
exudate from th~ living trees (wood and gum rosin). In chemical composition,
it comprises some 90 % of socalled resin acids and ca. 10 % non-acidic material
(e.g. esters of rosin and fatty acids, and various hydrocarbons). Rosin acid is a
N. A
C,H. B >
C
D
-5/10°C A N.
B VllC\Jum
C
D sample
Figure 3.1. Butadiene flow "cherne (Ihick line) for the distillation unit (left), dosage vessel (middle), wui polymeriuuion reactor (right). Appendages include pressure indicator (PI), pressure recorder (PR), and temperature recorder (TR).
~ 't:I ~
:I. i ~ 1:1 ... !!. "'= ~ .. c.. c: ~ '" ~ 1:1 c..
~ =r' 1:1 .E' = .. <I>
20 Chapter 3
complex mixture of monocarboxylic acids of alkylated hydrophenanthrene nudei,
the main components having the formula C19Hl9C02H and containing two
double bonds per molecule. It dissolvcs in strong alkaline solution~ to give rosin·
add soaps. Rosin acid soaps by themselves are unsatisfactory for use in emulsion
polymerizations, due to the presence of conjugated double bonds of the abietic"
type acids, giving rise to severe retardation.6,7,g) The double bonds can readily be
eliminated by disproportionation, a process where the rosin is heated over a
palladium catalyst giving simultaneous hydrogenation and dehydrogenation. The
resulting mixture has dehydro-, dihydro- and tetrahydroabietic acid as main
components, and contains less than 0.2 weight % abietic acid. Several qualities
are commercially available under the trade name dresinate.
The performance of rosin acid soaps is strongly pH-dependent, with an
optimum value of ca. 10 - 1 L For this reason a moderate excess of potassium
carbonate (pK = 10.2 at 2S"C) is included in recipe 1, and for reu.sOnS of
comparison maintained in recipe 2. This is not essential for SDS, as its solubility
and performance depend only weakly on pH; minor effects arise from changes
in micellar size.9) Although the prime function of K + INa + carbonate is to
maintain a conMant pH, its presence gives rise to secondary effects, originating
from the increase in ionic strength of the aqueous phase (chapter 5). It is further
realized that micelles of SDS may be non-spherical in systems with a high ionic
strength .. ·g· 10) However, no specific effects on reaction kinetics were found.
3.4 Initiator
An initiator serves as a source of radicals that initiate and maintain the free
radical polymerization process. AI elevated temperatures peroxodisulfate (as
ammonium, sodium or potassium Salt) l~ the most common initiator employed.
The decomposition of peroxodisulfate (hereafter pensulfate) in aqueous
media has been studied extensively. An excellent review ofliterature up to 1980
has been given by Behrmann and Edwards]])
Experimental Proeedutes Bnd Ted .... iques 21
In weak basic and neutral solutions decomposition occurs through a thermal
homolytic scission of the a-a bond, often considered a unimolecular reaction:
(3.1)
although there are strong indications that the homolytic scission may also occur
through a bimolecular reaction involving a water molecule:
Steps after the radical formation include:
(3.2)
(3.3)
(3.4)
Hydrogen peroxyde decomposes to form oxygen. In the presence of monomer
the radical intermediates are capable of initiating polymerization.
At pH <: 3, the following process contributes increasingly to the
decomposition of persulfate, without producing radicals:
(3.5)
Although such acidic solutions are of little significance to emulsion
polymerization, it emphasizes the necessity to indude a buffering substance in
the recipe to maintain pH > 3, as HS04- is formed in steps (3.2) and (3.3). This
is illustrated for the butadiene emulsion polymerization in section 3.5.
It has frequently been reported<·1f'2,13,14) that the presence of organic material
(including emulsifiers and monomers) increases the rate of disappearance of
persulfate, mainly due to oxidation of the material present in solution. Although
this does not increase the effective rate of radical production, the phenomenon
should not be ignored, as excessive consumption of persulfate leads to a
premature exhaustion of the initiator and incomplete polymerization.
22 Chapter 3
In order to check thi~ in the present system, the decomposition of persulfate
In the presence of dresinate 214 (recipe 1) was studied titrimetrically as a
function of temperature and emulsifier concentration.
Procedure~ the procedure followed is a modified version described by
Kolthoff et aL. IS) Latex samples collected at regular time intervals during
polymerization were cooled in ice under N~. The water content of the latex was
determined hy evaporating 10 mL of sample to complete dryness. About 20 - 30
mr., (containing ca. 0.2 mrnol persulfate) was weighed accurately in a centrifuge
tube, coagulated with 5 mL H2S04 (3 mo1.L'!), and centrifuged at 2500 rpm for
several minutes. The clear serum was transferred to a tItration vessel under Nz,
and the polymer rinsed thomughly with water. Potassium iodide (4.0 gram~) was
added, and after 30 minutes the iodine titrated with 0.01 moI.L'! sodium
thiosulfate; the end-point was determined with a starch solution. The potassium
persulfate concentration per unit volume of water [I] foUows readily from the
stoichiometry of the titration.
~: a plot of In [I] vs. time gave linear curves, where the slope yields
the first-order decomposition rate coefficient kd (Table 3.2). The presence of
dresinate 214 obviOUSly increases the rate of disappearance of persulfate hy a
factor of 1.5 - 2, in agreement with results reported for other emulsifiersy,)6) A~
thc emulsifier concentration C~ is decreased, k.J also decreases towards the value
found by Kolthoff and MilIerl 1) in a solution of 0.1 mol.L-1 NaOH. a strong
indication of a direct chemical interaction between persulfate and the emulsifier
or components therein. Yet, in this case such a side-reaction would hardly affect
the rate of radical production R; consequently the value for k~ of Kolthoff and
Miller (i.e. 6.3 10-<; 5-1 at 62°q was used to calculate R.
Polymerization times are of the order of the half life times of the initiator
(Table 3.2). Nevertheless, the significant decrease in persulfate concentration
does not appear to affect tbe steady-state kinetics, as interval II remains linear,
indicating an 'ideal' Smith-Ewart case 2 situation where Ii is independent of [I],
Of, as will be shown later, an extremely low initiator efficiency.
Experimental Procedures and Techniques 23
Table 3-2, Decomposition rate coefficients kd and half life times f1/2
for szol at various temperatures 0 and (;'mulsijier concentrations C"
8 CE lo".k<J t l / 2 Ht_k/l ·C L-I $-1 hr ~-1 g.
57 32.4 6.9 28 3.1 62 32,4 12.5 15 6.3 62 16.2 10.8 18 6.3 62 8.1 8.3 23 6.3 67 32,4 225 9 13_1
.) Data by Kolthof! and Miller l ?) in a solution of 0.1 mol.L-l NaOR
Other dissociative initiators used were 4,4'.azohis(4·granopemanoic acid)
(ACPA, Fluka AG, Buchs, Switzerland) and 2,2"azoisobutyronitrile (AIBN,
Fluka AG, Buchs, Switzerland). ACPA was uscd as received, whereas AIBN wa!S
recrystallized from methanol and stored at SOc. ACPA is completely water soluble at high pH, whereas AlBN will be
partitioned between the aqueous and organic phases present. Although AIBN
is often referred to as an oil soluble initiator, ig water sohlbility is appreciable,
about 13 mmol.L"l.lS) In fact, in the present study AlBN behaved similar as the
other, water·soluble initiators. The decomposition rate coefficients for ACPA
and AlBN in water at 0 '" 62"C were taken from literature: 1.2 10-5 g-I for
ACPA,19) and 6.3 1O~ 5-1 for AlBN.IS)
3.5 Buffer
In the preceding section 3.4 it was shown that when using persulfate as
initiator, the pH tends to decrease due to the formation of HS04', At pH < 3,
decomposition is predominantly via the acid.catalyzed pathway (equation 3.5),
without producing radicals capable of initiating polymerization_ It is thus
essential to maintain the aqueous solution neutml or basic, especially when
initiator efficiency with regard to the polymerization is low_ A~ a result of the
24 Chapter 3
long reaction times at relatively high polymerization temperatures, the emulsion
polymerization of butadiene is particularly susceptible to this phenomenon.
Figure 3.3 shows a pOlymerization with 50S as emulsifier (recipe 2), in the
absenl;c of sodium carbonate. A~ pH drops, polymerization becomes severely
retarded. A similar effect for the emulsion polymerization of butadiene using
sodium dodecyl benzenesulfonate as emulsifier was reported by Bhakuni. ZO) In
contrast, many other monOmers (e.g. styrene/I) vinyl acetate,22) various acrylatcs
and methacrylates23)) can readily be polymerized via emulsion polymerization up
to high conversions in the absence of a buffering substance.
#. 20 8 I ~
0. -0-0
(;;
\ 0 15 6 'iii ,....e
~ ...--e
0 ,....-<;; \ .,.... e 0 10 4 0 o ./
_"><0 Figure 3.3. Conversion (_) / ""'0 --0---'0-0_0 and pH (0) versus time for 5 • 2 a polymerization with 324 g.L"I SDS (recifJf;l 2) in the
0 0 absence of NazCD; .
0 2 4 6 8 24 Time in hours
),6 Chain transfer agent and gel analysis
Chain transfer agents are routinely used in emulsion polymerizations of
diene-monomers, to control the extent of branching and cross-linking of the
polymer. Thiols of low water solubility are mostly employed for this purpose.
In this study .evcral types of dodecanethiol were used: (1) an industrial
grade tertiary dodecanethiol (Pennwalt Chemicals, Rotterdam, Holland), which
is a crude mixture of CI2-isomers, containing small amounts of othe( compounds
Experimental Procedures and Tecbniques 25
(e.g. higher and lower thiols, isomeric alkanes), and (2) a well-defined mixture
of two isomeric tertiary dodecanethiOIS (Fluka AG, Buchs, Switzerland):
(CH3hCCH2C(CH3hCH~C(CH~)2SHand (CH3hCCH2C(CH;)(SH)CH2C( CH3h
Unless stated otherwise, tertiary dodecunethiol (t-DT) refers to the commercial
grade material supplied by PennwalL Furthermore, (3) n-Dodecanethiol (n-DT,
98 %, Merck, Darmstadt, FRG) was occasionally used in this study.
Although the gel content of a cro~s-linkable polymer is theoretically well
defined as the fraction of material of "infinite" molecular weight, experimental
criteria are usually more arbitrary, and strongly depend on the experimental
procedure. The same holds for the gel point, i.e. the conversion at which the
first insoluble polymer network makes its appearance. Therefore, the procedures
followed to determine the gel content and gel point in this study will be
described in SOme detail.
The first method is based on a toluene extr"ction of the soluble fraction of
the polybutadiene (PB)_ The insoluble residue is by definition the gel fraction_
The polybutadiene is isolated from the latex by precipitation in acidic methanol,
followed by repeated washing with water and methanol. The polymer is dried
under nitrogen at 500C. Toluene (100 mL) i~ pipetted to 1 gram of an accurately
weighed PB sample in a stoppered flask. After gentle shaking for at least 48
hours at rOOm temperature, the contents is passed over 11 100 mesh filter, and
50 mL of the filtrate evaporated to dryness in a weighed aluminium diSh.
The gel content is simply calculated according to (w •• - 2w")/w,,, where wr •
is the weight of the PH sample and wF is the weight of the dried filtrate. Linear
extrapolation to zero of the gel content as a function of conversion, gives the
experimental gel point This can be markedly different from the theoretical gel
point, since any possible gel stmctures of latex particle dimensions or smaller
(Le_ microgel) that persist after coagulation, remain undetected by this method
as they pass through the filter. It is assumed that during isolation of PB from the
latex the microgel agglomerates completely.
26 Chapter 3
The socalled viSlex methill.! is ba.~ed on measurements of the intrinsic
vi~cosity ['1] of diluted latex solutions_ The procedure is a slightly modified
version of the one described by Henderson and Legge?l) One gram of latex is
dissolved in 100 mL toluene"i50propanol (80 : 20 by volume), and diluted with
pure toluene, succe5sivdy. The viscosity was measured with an Ubbelohdc
viscometer at 25'C, and ['1) in pure toluene is obtained by extrapolation of the
reduced or inherent viscosity to zero concentration. A plot of (I)] versus
conversion has a pronounced maximum, taken as the gel point_
Initially ['II increases with conversion as molecular weight increases through
branching and cTOss·linking. Although the presence of toluene-swollen microgei
may stili raise ['1] somewhat, as the amount of microgel and the cross-link
density therein increa.~es, the contribution to [IJ] inevitably decreases and a
maximum in the ['7] vs. conversion curve occurs.
3.7 Particle size analysis
The average particle diameter was measured by dynamic light :;cattering
(DLS, Malvern Hc) and transmission electron microscopy (TEM, Philips 420 and
Jeul 2000 FX). For TEM the latexes were hardened with 0&04 and typically
some 750 - 1000 partkles wcre counted with a Zeiss TGA.l0 particle analyzer
From DLS only the weight·average diameter d" (calculated from the measured
z-average diameter) was used, while lEM gives complete information On
particle size and size distribution. Relevant definitions are given in Table 3.3.
The particle number density per unit volume of water (N) was calculated as:
N 6 c (M/W)
(ppi Paq) 1r d3 (3.6)
where c is the fractional conversion, (M/W) the monomer to water weight tatio,
P.q the water dcn~ity, Pp the polymer density, and d the particle diameter.
Experimental Procedures and Techniques
Table 3.3. Definitions of relevant pwticle size parameters, were n; represents the number of panicles with diameter d;
d. d 2 . d.} d,.. P
I:1l;.di Enj.djl !:1l;.di
3 Enj.dl4 d,.
Ell; Ell; Ell; EnlA3 d.
Subscripts: n denotes number-average, s surface-average, v volume-average, and w weight-average diameters.
27
Comparison between both methods of panicle sizing on the basis of d,. gives
excellent agreement; numerical values usually coincide within 5 % (Table 3.4).
However, the particle numbers calculated from these diameters are by definition
systematically smaller than the more correct ones based on the mean-volume
average diameter d,. In the present study this only affects numerical values by
some 10 - 20 % (e.g. Table 4.2 and Table 5.3), because the particle size
distributions are relatively narrow (average polydispersity P '" 1.08; standard
deviation = 0.03). The main conclusions are thus independent of the method of
particle sizing, which makes DLS, being faster and more convenient than TEM,
a very useful method.
Table 3.4. Comparison between average patti de diameters obtained with transmission electron microscopy and dynamic light scattering.
conversion (%)'l do/nm d,./nm
P TEM DLS
15 34 38 38 1.11 26 46 54 56 1.17 99 64 68 70 1.07 98 80 88 89 1.ll 27 100 113 116 1.13 41 130 139 139 1.07 57 152 161 162 1.07 96 183 190 191 1.04
a) Latex samples taken from different polymerizations (recipe 1).
28 C\1apter 3
3.8 POlybllladiene
The microstructure of polybutadiene (PB) consisting of sequences of cis-l.4,
trans-lA, and vinyl-l.2 units has been studied extensively with various
techniques.25,M,Z"I) For PB prepared by emulsion polymerization the fractions of
these microstructural units are statistically determined, and depend only on the
polymerization temperature. A cheek by Be NMR with a randomly chosen P8·
sample, using the triad determination method developed by Van der Velden et
al.lS) gave values in good agreement with other data (between brackets data
from Hampton29) using lR);18 % vinyl-1.2 (19 %), 20 % cis-1.4 (23 %) and 62
% trans-IA (58 %) units, at 8pol '" 62"C.
The microstructure (and the degree of cross-linking) may inOuence the
density of polybutadiene_ This was assessed experimentally by using butadiene
free latex samples taken at several conversions. The denSity of the latex was
measured with a calibrated Anton Paar precision density meter equipped with
a remote cell DMA 401, and the density of PB calculated assuming additivity
of the specific volumes of the aqueous phase and the polymer phase;
1 + (3.7)
where w is the weight fraction and p the density, while the subsripts 1, p and
.q denote the latex, polymer and aqueous phase, respectivdy. The density P.q
wa~ determined separately. From Table 3.5 Pp appears constant and independent
of conversion, viz. 0.878 g.mL-1 at 62"e. This value is in good agreement with the
onc determined by Mandelkern et al.30) at 25"C, viz. 0.892 g.mL-l
•
Table 35 Polymer density Pp detennined with a precision density meter, using lalex samples taken at several conversions c (.rtandard recipe n-
14
0.884
27
0.876
39
0.880
52 65
0.880 0.878
77
0_878
82
0.874
99 0.876
Experimental Procedures and Techniques 29
In an attempt to determine the saturation solubility of butadiene in
pOlybutadiene, the method outlined by Meehan3\) based on vapour pressure
measurements was tried. In short, a latex with an accurately known .,8 content
was equilibrated with liquid butadiene, while stirring constantly with a magnetic
stirring bar, The pressure was measured with a manometer, keeping sample and
manometer both immerged in a water bath at 2S"C. This wa~ done at various
butadienejPB ratios, while a correction was applied for the amount of butadiene
in the gas phase. A typical example is shown in Figure 3.4. Interpolation might
yield the saturation solubility of butadiene in PB, a situation equivalent to the
beginning of interval III. However, interpolation is not straightforward as
butadiene vapour does not obey Henry's law. The point where pressure reaches
the plateau value was taken instead.]l)
1.0
o
0.0 +----,---r--......... ----,--,-----r----, 0,0 0.5 1,0 1.5
Buladiene/PB in (g.g")
Figure 3.4. Vapour pressure at 2ye of latexe,f with varying butadiene! PH ratios. The lalex used stems from a polymerization with 32-4 g.L·
dresinale 214 (recip;: 1), having a conversion of 96 % (see Table 3.6),
Measurements were performed with ~amples taken at several conversions
(Table 3.6) during a polymeri:?ation with 32.4 g.L· j dresinate 214 (recipe 1). An
average Yalue of 0.95 gram butadiene per gram of polybutadiene was obtained,
30 Chapter 3
in agreement with data from Meehan31) (0.92 g.g.l) and Morton et al?~) (0.8 ±
0.05 g.g-I)_b) A value of 0,95 g.g_l would lead to a conversion value of 51 % for
the beginning of interval III, wheo all the butadiene is absorbed by the
maximum swollen latex particles.
Table 3.6. Solubility of buradiene in PB measured with latex samples taken at several conversions during a standard polymerization (recipe 1).
conversion (%) solubiJity (g,g-l)
26 0.98
.J ShQwn in Figu,e 3.4,
35 45 54 66 76 1.09 0.87 0.98 0.86 0.94
96')
0.92
Despite the relatively large experimental error, it appears that monomer
solubility is independent of conversion. However, the determination of the point
where pressure begins to drop (or alternatively a rough interpolation to yield
this point), is not neccessarily equal to the beginning of interval III. In Chapter
7 an unambiguous and exact method will be presented to calculate this poiot,
based on the polymerization kinetics in imervals II and Ill. According to thi~
method. the beginning of interval HI is at 60 % conversion, and was found
independent of particle size.
b) All data obtained by vapour pressure measurements.
References 31
1 Ger. 2,223,330 (1972), The International Syntheric Rubber Co. Ltd., Inv.: DA Brockenhurst, E,W, Duck; Chern. Abstr. 78, 59536t (1973)
2 US 4,385,157 (1983). Monsanto Co., Inv.: RM. Auclair, P. Rajendra; Chern, Abstr. 99, 23856s (1983)
3 US 4.783,508 (1988), Dow Chemical Co., Inv.: E.R Moore, TD. Traugott; Chern. Abstr. 110, 77261h
4 D.C. Blackley, "Emulsion Polymerismion - Theory and Pracrice", Applied Publishe,s Ltd., London 1975 , p. 291 • 300
5 J.B. Class, "Natural Resins~ in "Encyclopedia of Polymer SCience and Engineering", edited by H.F. Mark, N.M. Bikales, e.G. Overberger, G. Menges, Wiley-InteTSclence, New York 1985, 2nd edition, vol. 14
6 J.L. Azorlosa, Ind. eng. Chern. 41, 1626 (1949) 7 J. Stricker, Plaste Kautsch. 16,481 (1969) 8 J. Stricker, F. Fischer, W. Hoffmann, Plaste Kaulsch. 17,253 (1970) 9 A. Rahman, C.W. Brown, 1 Appl. Polym. Sci Z7, 2563 (1982)
10 D. Attwood, A.T. Florence, "Surjactants Sysrems", Chapman and Hall, London 1983, p. 80 - 85
11 EJ. Behrman, J.~. Edwards, Rev. [norg. Chem. 2, 179 (1980) 12 I.M. Kolthoff, 1.K. Miller, 1 Am. Chern. Soc. 73, 5118 (1951) 13 M.S. RyabQva, S.N. Sautin, N.L Smirnov, Zh. PrikL Khim. 50, 1719 (1977):
English transl.: J. Appl. Chem. USSR 50, 1648 (1977) 14 M.S. Ryabova, S.N. Sautin, NJ. Smirnov, ZIz. Prikl. Khim. 52,2065 (1979);
English trans!.: 1. AppI. Chem. USSR 52, 1951 (1979) 15 I.M. Kolthoff, LS. Guss, D.R. May, A.!. MedaJia, J. Po/ym. Sci. 1, 340
(1946) 16 C.E.M. Morris, A.G. Parts, Makromol. Chern. 119,212 (1968) 17 I.M. Kolthoff, lX. Miller,1 Am. Chem. Soc. 73,3055 (1951) 18 H. Bauer, M. Ortelt, B. Joos, W. Funke, Ma.kromol. Chem. 189,409 (1988) 19 D.C. Blackley, A.C. Haynes, J. Chern. Soc., Faraday Trans. I 75, 935 (1979) 20 R.S. Bhakuni, Ph. D. ThesLr, University of Akron (1964); Chern. Abstr. 62,
13351b (1965) 21 D.C. Blackley, S. Andries, R.D. Sebastian, Brit. Polym. J. 21,313 (1989) 22 A. Penlidis, J.F. MacGregor, A.E. Hamielec, J. Appl Polym. Sci. 35, 2023
(1988) 23 N. Sutterlin, RJ. Kurth, G. Markert, MakromoL Chern. 177, 1549 (1976) 24 D.A. Henderson, N.R. legge, Can. 1 Res. B27, 666 (1949) 25 S.W. Cornell, J.L. Koenig, Macromolecules 2, 541 (1969) 26 R.J. Harwood, Rubber Chern. TechnoL 55,769 (1982) 27 H. Sato, K. Takebayashi, Y. Tanaka, Macromolecules 20, 2418 (1987) 28 G. van der Velden, C. Didden, T. Veermans, J. Beulen, Macromolecules
lO, 1252 (1987) 29 RR. Hampton, Anal. Chern. 21, 923 (1949) 30 L. Mandelkern, M. Tyron, FA Quinn, 1 Polym. Sci. 19,77 (1956) 31 E.J. Meehan, 1. Am. Chern. Soc. 71, 628 (1949) 32 M. Norton, P.P. Sa.latieIlo, H. Landfield,1. Polym. Sci. 8, 215 (1952)
POlymerizations with Rosin Acid Soap 33
Chapter 4 Polymerizations with Rosin Acid Soap
SUMMARY: The kinetics of the emulsion polyme,i:wtion of butadiene was investigated, using dresinate 214 as emulsifier in combination with three dissociative initiators, namely potassium persulfate. 4,4··azobis(4-cyanopentanoic acid) and 2,2'-azoisobutymnitrile. The reaction rate Rp<>I in interval n was found to be hi~hly insensitive to changes in the initiator concentration (~I Ol [I] .00), indicative of a low initiator efficiency. The development of particle number N as a function of conversion at several emulsifier concentrations [E], reveals a limited colloidal stability of the particles. Rp<>l depends on [E] with an exponent of 0.61, whereas N after cessation of coagulation has an exponential dependence of 1.6. As a consequence R,.ctlN (and thus n) must depend on particle size, since the monomer concentration within the particles is virtually constant in interval n. A certain analogy in behaviour between the emulsion polymerization of various polar monomers, kinetically dominated by radical desorption, and the emulsion polymerization of butadiene, suggests that similar events determine the kinetic course in the present system.
4.1 Introduction
Despite the industrial importance of the emulsion (co)polymerization of
butadiene, very little has been reported about its kinetic and mechanistic
features. The aim of this research is to investigate the effects of various reaction
parameters on the kinetics of the emulsion polymerization of butadiene,
deliberately sta,ting from a typical industrial recipe. This chapter describes (he
influence of the initiator and emulsifier concentration, using disproportionated
34 Chapter 4
rosin acid ~(lap as emulsifier (recipe 1), and ,cveral dissociative initiators
differing in ~trueture and water solubility. All experiments were conducted in the
presence of a thinl as chain tran~fcr agent, in concordance with general practice.
4.2 Effect of the initiator- concentration
In earlier investigation, it was found that the rate of emulsion
(co)polyrnerlzation of butadiene(75)-styrene(25) or butadiene alone (using the
CR-S recipe) remained virtually constant even when the concentration of
persulfate wa~ changed lOO-fold. l ) In order to get more insight into this
phenomenon polymerizations were performed with various types of dissociative
initiat()r~: potassium persulfate (PPS), 4,4'-azobis(4-cyanopentanoic acid)
(ACPA) and 2,2'-azoisobutyronitrile (All3N). The first two initiators are
completely water-soluble unocr the experimental conditions (pH'" 10.5 - 10.8),
whereas AWN is only sparingly soluble in water. The overall [K·) was kept
constant at 0.4 moI.L·' by appropriate adjustment of the amount of K2C03.
~3,0 '",
~ E 2.0 0)
:::::: 8.
0::
"b ~ 1.0
0.1 0,3 0.6 1,0 3.0 6,0 10 30 1 O·,3. R/(g·' .s·')
Figure 4.1, Variation oj the overall polymerization rate Rpc' witfJ the radical production rate per gram emulsion R Jor three different iniliators: PPS (0),
ACPA (t:.), ilnd A/BN (7), l.l.5ing dresinate 214 as emulsifia (recipe 1).
Polymerizations with Rosin Acid Soap 35
Figure 4.1 shows the effect of variation of initiator concentration on
pOlymerization rate Rpo'. The rate of radical production per gram emulsion R
is calculated according to R = 2 kd [I) N ... " where kd is the decomposition rate
coefficient, [I] the initiator concentration in moles per gram of emulsion and N"v
Avogadro's number. Values for ~ at polymerization temperature (62'C) were
calculated from data in the literature: PPS2) 6.3 100(\ 5"1, ACPAl ) 1.2 10-5 S-I and
AIEN 1.3 1O-~ S-1 in toluene,4) and 6.3 100(\ S-l in water.S) As all values fOr ko are
of comparable magnitude, the net effect on R will be small.
It appears that the polymerization kinetics is highly insensitive to the
initiator concentration, irrespective of the nature of the initiator (Rpol oc (Ito()!\ Inspection of the conversion-time curves (Figure 4.2) indicates that particle
formation is a slow process since the nucleation period (ioe. interval I) is
unusually long, typically some two hours or more. The total number of primary
radicals generated during this period is 1017 - 1018 per mL of water, while the
particle number is of the order of lOIS per mL of water (see Table 4.2)0 This
large discrepancy implies that all three initiators investigated are very inefficient
in the particle nucleation process. Therefore, the relatively small variation of
initiator concentrations jn the studied region will hardly affect the kinetics.
If? 100 J::
J:: 0 SO 0to")
~ t;; 0 60 0
40
20
0
0 Time in hours
Figure 4-2. Conversion-time curves of persulfale-irlitiated polymerizatiofL~ at different initiato,. concetUraJions (I] '" 13.0 mmolL1 (fj.); {If " 2.6 mmo/.L-l (0); {II'" 0.52 mmol.L"1 ('7); and {II'" 0.26 mmol.Lol (0)-
36 Chapter 4
The low initial rates and long duration of inte.-val I cannot be attributed to
the presence of any possible retarding components in the dresinate 214 (e.g.
ahietic acid-type of derivatives, section 3.3), as similar results werc obtained with
other emulsifiers, including sodium dodecylsulfate (see chapter 5).
A comparison between the experimental values for the particle number N
and the critical time te, (the duration of interval 1) On the one hand, and the
predictions by the Gardon theory6) (being an extension of the Smith-Ewart
thenry) in Table 4.1 shows that the prcsent system deviates significantly from the
'iocal' emulsion polymeriz;\tion as embodied in the Smith-Ewart assumptions.?)
It should be noted that Gardon's treatment is only applicable to cases with
negligihle raJical desorption (n ::: O.S).
The nature and ionic charge of the primary raOieais evidently have no e;ffect
on the reaction kindies, indicating that the reactivity of the oligomeric radical
species initiating polymerization in the loci is dominated by the hydrophobic
moiety. Similar effects have becn reported for the emulsion polymerizati(m of
styrene, using initiator systems generating anionic, neutral or cationic radkals.S)
Table 4.1. Experim(:lzlal arid theoretical values for the particle number N, and critical time tor (see text) for polymerizations with different initiators.
Initiator e in % dJnm 1O-15.N/mL·l lO->'t«/s
expo Gardon') exp.b) Gardon')
PPS 95 83 1.S 10 9.5 0.18 PPS 98 87 1.4 10 9.5 0.18
ACPA 99 88 1.3 13 8.5 0.13 AlON 94 84 1.5 10 - 14') 9.0 0.12 - 0.18<)
0) Calculated using a value of 100 Lmor1.s-1 for the propagation rate c()effi~ient at 62<>C,9) and an initiator concentration (I] =- 13 mmoJ.I .-1.
n) Roughly estimated from the conversion-time curves as the elapseJ time 11lI polymerization rate becomes constant.
0) Calculated with the values for k~ in toluene and water.
Polymerizations with Rosin Acid Soap 37
4.3 Effed of the emulsifier roncentration
Figure 4.3 shows the conversion-time CUrves of polymerizations in which the
emulsifier concentration was varied eightfold between 64.8 g.L· j and 8.1 g.L'I.
All experiments were conducted above the critical micelle concentration (CMC)
of dresinate 214, determined tensiometrically with the DuNouy ring method to
amount to 1,2 g,L·1, i.e. about 3.5 10-3 mol.L-1 (0 "" 25°C and [KzC031 .. 0.15
mol.L"I). Agreement with values reported in literature is good if differences in
electrolyte concentration are taken into account: 10.4 10.3 mol.L"1 for dresinate
21410) and < 10-3 mol.L"1 for "sodium rosinate" (presumably dre~ina.te 731),1)
both values determined at 50·e and without additional electrolyte.
rf!. 100 I::
I:: 0 80 .~
~ t:: 0 60 ()
40
20
0
0 2 4 6 B 10 12 14
Time in hours
Figure 4.3. Conversion· rime Curv.;:.)" of persuljale-initiated polymerizations at different dresinate 214 concentrations Co '" 64.8 g.L.I (l!.); Co "" 32,4 g.L'!
(0); C. = 16.2 g.L·1 ('V); and CIS '" 8-J gLI (0).
Plotting the polymerization rate in interval II again~t the emulsifier
concentration on a log-log scale yields a value 0.61 for the emulsifier exponent
(Figure 4.4), This is in excellent accordance with the Smith-Ewart case 2
38 Chapter 4
behaviour, insofar as Rpol i, proportional to N, However, in the prcsent system
lhi~ agreement is apparent, as will be explained in the following.
Data on particle si:1:e and number as a function of conversion are given in
Table 4.2 and Figure 45 (see page 40), Agreement between TEM and DLS is
excellent, Within each experiment the particle number reaches a constant value
at conversions greate, th,in ca. 40 %, however, the behaviour at lower
conversions is of more jntere~L In the intermediate emulsifier concentration
range:: studied, a steady decrease in particle number is clearly observable. This
implies that beside particle nucleation (micellar or homogeneous) a second
mech,mism is operative which is coagulative in naturC. This process is relatively
slow since it could be detected using conventional sampling procedures.
~ 4.0
E • --.....
II: ~ " 2.0 b 40
'''1 ~
1.0 2.0 Ei /i/ £! --0.6 3. rr: 1.0 'b
04 ,/; Figure 4.4, Dependence of 01' Rpo[ (6) and final N on 0.2
04 emulsifierconcentration C£ Particle.1; sized with TEM (e) and DLS (0). 0.1
e 10 20 40 60 100
C,:/(g.L· I}
Evidence is accumulating that coagulation of latex particles cannot be
neglected in emulsion polymerization. II) It has been shown that limited
coagulation determines the particle number in the emulsion polymerization of
more water-soluble monomers (e.g, vinyl a<:etate,ll) ethyl a<:rylate,l3) methyl
Polymerizations with Rosin Acid Soap 39
methacrylateI4)), and of non-polar monomers in emulsifier-free syStems. IS)
Recently Feeney, Napper and Oilbertl6) proposed a coagulative nucleation
mechanism combining MUIIer-SmoJuchowski coagulation kinetics with the DLVO
theory of colloidal stability;'g. 11) thus suggesting that limited coagulation is
inherent in any emulsion polymerization.
Table 4.2. Particle diameter d and particle number N detelmined with TEM and DLS cU different concentrations of dresinate 214 (recipe I).
CJg.L-1 cin % dv!nm dw/nm Hy!s.N/mL I .)
TEM DLS TEM DLS
64.8 74 55 59 61 4.1 3.2 99 65 68 69 3.3 2.9
32.4 15 35 37 38 3.2 2.6 26 49 54 56 2.1 1.6 41 6S 71 73 1.4 1.1 68 83 88 89 1.1 0.92 98 87 93 97 1.4 l.l
16.2 14 42 43 45 1.7 1.6 26 67 n 71 0.79 0.65 40 94 100 104 OA5 0.37 70 126 132 134 0.32 0.28 99 143 149 153 0.31 0.28
8.1 13 78 90 90 0.25 0.17 27 104 113 U6 0.22 0.17 41 134 l39 139 0.15 0.14 57 155 161 162 0.14 0.13 97 185 190 191 0.14 0.13
a) N based on dv when using TEM.
It is obvious that coagulation will influence the initiator efficiency toward
particle nucleation, the duration of interval I, and the final particle number.
Deducing the emulsifier exponent from a linear log-log plot of final particle
number verSus emulsifier concentration (Figure 4.4), gives a value of 1.6. The
corresponding exponent toward polymerization rate Rpot was found to be 0.61.
Since the butadiene concentration in the latex particles was found to be
practically constant in interval II and independent of emulsifier concentration
40 Chapter 4
(chapter 7 and references therein), the average rate per partide Rpol/N must be
a function of particle size to account for the observed behaviour.
Similar hut unexplained results for the homopolymerization of butadiene
have been rcported by Morton et al. IB) and Wendler et al.. 19) The latter found
a distinct effect of the amount of bis(isopwpoxythiocarbonyl)disulfane on
particle size, as a result of changing the colloidal properties of the system.2(l)
Experimental results are summarized in Figure 4.6 were d9(l represents the
particle diameter at 90 % conversion, calculated from the experimental data
assuming a constant particle number: d9(l = (0.9 d3/e)l/3.
:... 3.5
E .. ---::!'; 3.0 .. b
"' __ .-.:"':...... __ .. __ "_A __ ~
2.5
2.0
1.5
1,0
05
0,0
0 20 40 60 80 100 Conversion in %
Figure 4.5. Van·alion of particle number N with converSion for polymerizations with different dresinate 214 concentratiolLf C~ (symbols see legend of Figure 4.3). Full symbols indicate TEM data, and open symbols DLS data (both using dJ.
Using Rpo/N as a semi-quantitative equivalent of the average number of
radicals per particle n, it becomes obvious that the larger particle~ will have
higher values for n and thus will grow more rapidly. For polydisperse latexe~ this
means that the poJydispcrsity increases upon polymerization, an effect that is
enhanced hy the coagulation of pa{tides. Figure 4.7 shows cumulative particle
Polymerizations with Rosin Acid Soap
~60 '(I)
.8l40 -., Z i
IIi 20 'b
10
6,0
4,0
2.0
'.0
0.6
20
/ 'i!o a/ It II ~ ~
/~ D
I
40 60 100 200
dw/nm
41
Figure 4.6. Average rate per particle (Rpo/NJ vs" panicle diametefilt 90 % conversion (d-xJ fordifferent butadiene emulsion polymerizations: experimental results at 62"C, lJ.$ing TEM (.) and DLS (0) for particle size analysis. D(lIa of rq:J9) at ~pol ,. 7(J'C using dv obtained with TEM (0); data of ref. 18)
based on suiftu:e-trverage diameters d. (determined by soap titration) at 4(fC (v), 5crc (,'If) and 6fJ'C (6).
size distributions fot a single experiment at various degrees of conversion as a
function of particle si:;:e and volume. A positive skewness develops at low
conversion which becomes more pronounced as polymerization proceeds.
Surprisingly, the mechanistic complexity is not reflected in the conversion
time curves, which are simply S-shaped with a linear interval It A constancy of
rate is usually taken to imply that the particle number i, also constant, which is
definitely not true for the present system (Figure 4.8), In this particular case a
decrease in particle number through coagulation will practically be compensated
by an irtcrease in ii of the resulting particles, since R.,.,/N '" dl.8 a~ calculated
from the data in Figure 4.6. Other factors cannot be completely ruled out. For
example, the cross-linking behaviour of the polymer and its effect on monomer
42
concentration and rate coefficients may playa role. If present, their net effect
on polymerization rate i~ very small and remained undetected, because Rp>1 was
fOlJnd constant within experimental error. '
An analogou~ behaviour is known to occur if more water-soluble monomers
are used: 12,13,14) a decreasing particle number while,the rate of polymerization is
constant. Radical deSIJrption is relatively facile with such monomers and Ii may
be much less than 0.5: A Smith-Ewart ca.~e 1 situation arises and RI"'I will only
depend weakly on particle number. This type of behaviour may also be found
with styrene a! .certain emulsifier concentrations.21 ) Dunn and Chongl:!) showed
that the vari,a!ion 9f partiCle number in the emulsion polymerization of vinyl
acetate is in accord with expectations on the basis of DLVO theory.
!!,100
0 0 80 ~
'" = ~
" 60
~ '3 ~
40
0
20
o 20 40 60 100 120 140
ctJnm
"" 100 r /Q,....--()?____''CI'..-1I' ~IJ ______ O--O
~
" i!! /0/0 .~ 00 Figure 4. 7. Cumulative panicle .-g if! / size distribution vs. particle ~
" so diameter- d. (lOp) and particle
l volume v (bottom) for-a S20/
~ 4() III initialed pol}!merizatinn ([IJ '"
8 13 mmolL-1, C. '" 32.4 g.el
,
:ill / I recipe 1) at various degrees of conversion: 15 % (A), 26 %
0 ,./ (0), 4/ % ('\1) and 98 % (0)-i I I
10 12 14 10·~.v/nm:'
Polymerizations with Rosin Add Soap 43
~ 100 .~ c o
. ~
~ 80 o u
60
40
20
..
pl"
" P
\ ,/'
o'
/' }'
o /
" " l \" ol
.. ,j \/
2.0 :..., E
"'""-:;;:;
0.0 ~
1.5
1.0
;:/1\"\,, 0.5 Figure 4.8. Conversion (0) • and panicle nu.mber vs. lime
of' o"--.," __ :_,,--1f- curves for a polymerization .0' with J6.2gL dresinate 214 .
. ero Particle number N based on o---¥o,'--o---,---.,-_-.---_.--'---T_ ....... ~t_+_oo d" (A), and on d .. (r",).
o 4 8 12 24 Time in hours
It is COmmOn practice in diene-polymerizations to use chain transfer agellts
such as thiols and bis(alkyloxythiocarbonyl)disulfanes to control the extent of
cross-linking of the polymer. Chain transfer agents were also found2Z.23) to
promote desorption of radicals from the latex particles and m; a consequence to
lower Ii. However, Nomura22) showed that n·dodecanethiol did not affect either
the rate of polymerization or the particle number in the emulsion polymerization
of styrene. lIe concluded that the CH~(CH.)l1S· radicals cannot desorb because
of their extremely low water solubility. It is not yet clear if desorption of tert
dodecanethiol radicals is significant in the present system.
Furthermore, small amounts of thiols of low water solubility seemed to be
essential to bring about reaction at ac~eptable rates in the persulfate-initiated
polymerizations of butadiene(75)-styrene(25) or butadiene.:IA) These important
effects of tbiob in the emulsion polymerization of butadiene were investigated.
and will be dis~ussed in chapter 8.
44 Chapter 4
Accurate caklllation of the absolute value of Ii i~ not yet possible due to the
large uncertainty in the value for the propagation rate coefficient l), cited in
literature,9) and the unknown distribution of Ii among particles of different sizes.
Based on the present experimental data no conclu~ive evidence can be provided
as to which mechanisms determine the reaction kinetics. Either radical exit or
first-order termination (e.g. by radical trapping) may suppress n. Moreover, slow
termination of radicals in the particles, which would raise Ii significantly, calUlot
be ruled out. The observed similarity in behaviour between the emulsion
polymerizations of certain polar monomers and butadiene, suggests a Smith
Ewart case 1 situation for the latter monomer as welL Nonetheless, it is
r;urprising that the observed anomalies are so pronounced for such a sparingly
water-soluble monomer in the size range studied (dso ;; 50 - 150 nm), while
emulsffier concentrations are well above the critical micelle concentration.
This initial work ra.ised many new questions: the low initiator efficienc)', the
ease of limited coagulation of PB-Iatexes, the size-dependence of ii and many
others_ Nevertheless, it became obvious that regardless of apparent simplicity of
the conversion-time history and partial agreement of experimental data with
Smith·Ewart theory, the mechani,m and kinetics of this polymerization are more
complicated than is usually assumed.
Polymerizations with Rosin Acid SQap 4S
4.4 Conclusions
The concentration of three dissociative initiators differing in structure and
water solubility described in this chapter hardly affects the reaction kinetics of
the emulsion polymeri:!;ation of butadiene. The number of primary radicals
necessary for nucleation exceeds the final particle number by at least two orders
of magnitude, so that the kinetics will become almost independent of the rate
of initiator decomposition. The nature and ionic charge of the primary radicals
also have nO influence on the polymerization, indicating that the hydrophobic
moiety of the oligomeric radicals originating in the aqueous phase governs the
initiation kinetics of polymerization in the loci.
The particle number is partially determined by limited coagulation of
colloidally unstable particles. Experimental results may not be compared with
theories (e"g" Smith-Ewart and Gardon) that neglect this phenomenon. The
particle number after cessation of coagulation (conversion> 40 %) varies with
the emulsifier concentration Ca to the 1.6th power, and the polymerization rate
in interval II with CR to the 0.615t power. This implies that the average number
of radicals per particle Ii is a function of particle size, since the monOmer
concentration within the latex particles is approximately constant in interval II.
Larger particles have higher values for ii and will grow more rapidly, Data
on particle size distributions are in agreement with this observation: a positive
skewness develops at low conversion, and becomes morfl pronounced as the
polymerization proceeds"
The kinetic similarity of the emulsion polymerization of butadiene with
those of certain more water-soluble monomers (e.g. vinyl acetate, various
acrylates) reported in literature is striking: the rate of polymerization is constant
whereas the particle number is decreasing. It has been shown that for the latter
emulsion polymerizations radical desorption is dominating the reaction kinetics,
leading to ii « 0.5, suggesting a similar situation for butadiene.
46 References
FA. Bovey, I.M. KOlthoff, A.I. Medalia. RJ. Meehan, "Emulsion Polymen..ation", Inlerscience Publishers, New York 1955
2 I.M. Kalthoff, LK. Miller, 1. Am, Chern, Soc. 73, 3055 (1951) 3 D.C. Blackley, A.e. Haynes, l. Chern. Soc., Faraday Trans. 1 75, 935
(1988) 4 M. Talat-Erben, S. Bywater, 1. Am. Chern. Soc. 77, 3712 (1955) 5 R Bauer, M. Ortelt, R Joos, W, Funke, Makromol. Chern. l89, 409
(1988) 6 J.L. Cardon, 1. Polyrn. Sci., Part A-J 6, 623 (1968) 7 W.V. Smith, R.H. Ewart,l. Chem. Phys. 16, 592 (1948) 8 LA. Penboss, D.H. Napper, R.C, Gilbert, J. Chem. Soc., Farcutay Trans.
1 79, 1257 (1983) 9 M. Morton, P.P. Salaliello, II. Landfield, J. Polym. Sci. 8, Z15 (1952)
10 S.H. Maron. M.E. Elder, I.N. Ulevitch,1. Colloid Sci. 9, 382 (1954) II A.S, Dunn, Makromol. Chem., SuppL 10/n, 1 (1985) 12 A.S. Dunn. L.C.H. Chong, Brit, Polym. I. 2, 49 (1970) 13 V.1. Yeliseyeva, "Polymerization of Polar MOhomers", in: ''Emulsion
PolymetUation", edited by l. Piirrna. Academic Press, New York 1982 14 R.M. Fitch, C.H. Tsai, "Polym(;!r Colloids", edited by R.M. Fitch, Plenum,
New York 1971, p. 73 15 AR. Goodall, M.e. Wilkinson, J. Heam,J. Polym. Sci., Polym. Chern. Ed.
15, 2193 (1977) 16 P.J. Feeney, D.H. Napper, KG. Gilbert. Macromolecules 17,2520 (1984) 17 R.I1. Oltewill, 'Tile Stability and Inst(1.hiiity of Polymer Latices", in:
"Emulsion Polymerization", edited by I. Piirma, Academic Press, New York 1982, p. 8
18 M. Morton, P.P. Salatiello, H. Landfietd,1. Polyrn. Sci 8, 111 (1952) 19 K. Wendler, N. Karim, M. Fedlke. Plaste Kaursch 30, 247 (1983) 20 K. Wendler. H. Elsner, W.o. Hergeth, M. Fedtke, Plasre Kav.tsch 32, 128
(1985) 2J S.l. Kincaid, Ph. D- Thesis, University of Akron (1983); Chern. Abstr. 99,
54209n (1983) 22 M. Nomura. Y. Minamino, K. Fujita, M. Harada, 1. Polym. Sei., Polym.
Chem. Ed. 20, 1261 (1982) 23 G. Lichti, D.F. Sangster, B.C.Y. Whang, D.H. Napper, R.G. Gilbert, 1.
Chern. Soc., Faraday Trans. 1 78, 2129 (1982) 24 LM. Ko\thoff, WE. Harris, J Polym. Sci. 2,41 (1947)
Polymerizations with Sodium Dodecylsulrate 47
Chapter 5 Polymerizations with Sodium Dodecylsulfate
SUMMARY: The emulsion polymerization of butadiene was investigated with sodium dodecylsulfate as emulsifier, and results compared with those obtained with the industrial grade emulsifier dresinate 214. Limited coagulation was observed in both cases, caused by the higb ionic strength of the aqueous phase. Sodium dodecylsulfate was found a better stabilizer than dresinate 214. The average rate per particle in interval II depends strongly on particle size, but not on recipe parameters used to vary the final particle diameter. The dependence of particle number N on initiator ([I)) and emulsifier concentration ([E]) is usually expressed as the exponent in the empirical relationship N '" [IJ'.[EF. Promoting coagulation by increa~ing the ionic strength, decreases x toward zero, while y increases soal"PJy. lncreasing [l] at a constant [E] and ionic strength induces limited coagulation, presumably because the rate of emulsifier adsorption becomes insufficient.
S.t Introduction
In the previous chapter the kinetics of the emulsion polymerization of
butadiene with the commercial emulsifier dresinate 214 was discussed, using a
thiol as chain transfer agent as customary in diene-polymeri~ations. Several
remarkable results were obtained with this 'industrial' recipe: the initiator
efficiency with regard to particle formation is low and independent of the type
of initiator, limited coagulation is involved in the particle formation process, and
no Smith-Ewart case 2 regime with n = 0.5 was observed. At that point it was
unclear if these phenomena resu1.ted from specific reaction conditions, such as
48 Chapter 5
the choice of emulsifier and the presence of a chain transfer agent. In order to
answer these questions, additional experiments were performed with the well
defined emulsifier sodium dodecylsulfate, under comparable conditions.
5.2 Effect of the emulsifier concentration
The concentration oC sodium dodecylsulfate (SOS) was varied between 64.8
and 4.0 g.L' (224 and 14 mmol,Ll), all concentrations well above the CMC,
which was determined with the Wi/helmy plate method to amount 10 0.17 g.L'1,
i.e, 0.6 mmoLL"' (0 "" 25'C and [N~C03] .. 0.15 mol.L·1), The overall [Na+] is
0.3 mol.l;\ and kept constant by adjustment of the concentration of NazC03•
The small variation in pH that may result is unimportant, since the performance
of SDS is virtually pH-independent at pH > 7, this in contrast to dresinate 2l4.
The [Na+) is thus an additional adjustable parameter in recipe 2 (section 53).
f12. 100 c c g 80
~ t;:
8 60
40
20
o 2 4 6 8 10 12 14 Time in hours
Figure 5.1. Conversion V$. time curves for po/ymerization.r (recipe 2) with differe~~ sodi~m d:!-ecyfsu/f~~e co~cent~ions C"-l "" ~.8 g.~'1 (lJ.); ~f ""
32.4 g,L (e), CE - 16.2 g.L (v), C£ - 8.1 c·L (.), CE - 4.0 g.L (0).
Polymemations wilh Sodium Dodecylsulfate 49
if 100 50 ~
c ~ <:; -0 z .~
i::J '" :> SO ;6°.0 4,0 c
0 0
/' " I
60 l
! :3,0
\ I I
40 /' 2.0 Q
I l . '"
Figure 5.2 COhversiQn (0) ... 1 '" I
and panicle number vs. time 20 p\ .. 1.0
curves for a F:.lymerizaJion ," . .,.Q , ...
with 8.1 g.L" SDS. Particle ,0.0 4-.........a-~-o.-t"""'*'-number N based Oh dy (.),
0.0 ,0'
0'"
and on d", (Ll). 0 . .,.
0.0 0 4 8 12 24
Time i~ hourS
All conversion-time curves (Figure 5.1) are convex toward the time axis up
to about 40 % c,anversion, this in contrast to similar polymerizations with
dresinate 214 (Figure 4.3), whe,e linearity already was attained at about 25 %
conversion. J~te\"VaI III begins at 60 % conversion (see chapter 7), so that
interval II appears rather short in the present system.
Figure 5.2 gives the particle number and c/?nversion vs. time curves for the
polymerization with 8.1 g.L·t SDS. Evidently, as in polymeriz;ation~ with dresinate
214 (Figure 4.8), limited coagulation is occurring with SDS as emulsifier. Also
in concordaoce with previous observations, the decrease of N by coagulation is
not showing up in the conversion. time curve, Le. 1\01 is not proportional to N,
so that ii cannot be constant at 0.5 (Smith-Ewart case 2). Indeed, N decreases
while RI"'I still increases. ~ is usually eJlpressed as:
(2.2)
50 Chapter 5
where k., is the propagation rate coefficient, ii the average number of radicals
per particle, and e" the monomer concentration within the latex particles.
The average rate per particle in interval II (R",,/N) can be used as an
equivalent for n, since the other parameters (Le. \ and eM) are constant in
interval U to within a good approximation (see chapter 7). Obviously, Ii must
depend on particle size to accoLlnt for the observed behaviour (figure 53). This
effect becomes ~trongly pronounced by the coagulation process, causing an
enhanced variation in particle diameter within one single experiment.
~
.::: 20
z 3.
0;: 10 '0
6.0
4.0
2.0
1.0
40 60 100 Diameter in nm
Figure 5.3. Average rate per particle RJN V.I'. putticle diamerer ,C(e) ami dw (0), for rhe pofymcriztllion with &1 gLi SDS from Figure 5.2.
The dependence of particle number (or rate) on emulsifier ([ED and
initiator concentration ([I]) is usually expressed as the exponent in the empirical
relatinnship N '" [I)".[E)Y. The theoretical values x,ft = 0.40 and Yth '" 0.60 are
obtained with any nucleation mechanism (homogeneous') or micellar) that
ignores coagulation. Experimentally obtained values for y (and y') in the
emulsion polymerization of butadiene are given in Table 5.1.
Polymerizations with Sodium Dodecyls.ulfate 51
Table 5.l. Emulsifier exponents with regard to N and Rpol
Emulsifier yin N ()[ [EJY y' in Rpol Q( lEy' Ref.
SDS 2.1 024 ch.5 Dres.214 1.6 0.61 ch.4 Oresinate 1.0 0.62 3)
Given the non· ideal kinetic behaviour of the present system (Le. Rpol ~ N),
the different exponents for 1\01 and N are not surpri5ing, while the significantly
higher values with regard to N as compared with theory (Y'h = 0.60), clearly
result from the coagulation process. The purity and structure of the emulsifier
can only be partially responsible for this; other parameters are also involved.
5.3 Effect of the sodium ion concentration
The general principles of electrostatic stabilization of colloidal particles are
well established in DLYO theory!·j;· 4) In short, electrostatic repulsion between
particles results from the presence of charged surface groups (from initiator
fragments or functional (co)monomers). or adsorbed ionic surface active agents.
As a consequence, the surface acquires a surface potential ¢o' giving rise to a
BoltzmalUl distribution of counter ions and co-ions in a region close to the
surface_ According to a simplified model for a spherical particle of diameter d,
the potential at a distance r from the centre of the particle 1/1, is ("'0 < 25 mY):
d ¢, '" lPo 2( exp[x;(0.5d - r)] (5.1)
The parameter II; is related to the ionic strength a by:
(5.2)
52 Chapter 5
where NM is Avogadro's number, e the fundamental electronic charge, €, the
relative pcrmittivity of the aqueous phase and "0 that of free space, k the
Soltzmann constant and T the absolute temperature. It is I< that mainly
determines the fall off of the electrostatic potential with distance from the
surface, and consequently the range of electrostatic interaction between particles.
From this simplified de~cription, it already becomes obvious that the ionic
strength is an important parameter in achieving colloidal stability.
The significance of ionic strength on colloidal stability of polystyrene
particles prepared by emulsifier·free emulsion polymerization has been
demonstrated by Goodwin et al..5.6) Limited coagulation is well e5tablished in the
emulsion polymeriz-ation of mOre water-soluble monomers/'S,?) caused by slow
and lor weak adsorption of emulsifier on the polar polymer-water interface as
suggested by Yeliseyeva.9) Dunn t) showed for vinyl acetate that the coagulation
observed experimentally can be described satisfactorily with DL VO theory.
Furthermore, the role of coagulation in the particle nucleation mechanism has
been treated quantitatively by several workers!&- 10,11,12)
However, butadiene is only sparingly water-soluble (37 mmol.L-1 at 50"C),13)
the emulsifier concentrations are far above the CMC, while coagulation extends
well beyond the nucleation stage and involves particles of 'normal' size.
As pointed out earlier, the [Na 'j in recipe 2 is an extra adjustable
parameter, which can be varied by changing the amount of sodium carbonate,
present in reasonable excess_ In this way we can separate the effects of ionic
strength on particle nucleation and growth phenomena, from those associated
with the emubifier and initiator, compounds that beside their prime function in
emulsion polymerization systems, also act as inert electrolytes that contribute to
the ionic strength 0, However, the anion concentration (and tbus cr) changes
continuously during polymerization, owing to the decomposition of the initiator
anu adwrplion of surface activc ions on newly formed surfaces. Therefore we
have chosen the overall cation concentration as [he equivalent of 0 (although
not exactly identical, since 1 : 2 electrolytes were used).
Polymerizations with Sodium Dode()hulfate 53
The overall rNa +) was varied at two emulsifier concentrations, viz. C., '" 32.4
g.L· ' and C. '" 8.1 g.L·I, while [I) was kept constant at 13.0 mmOILI. The final
particle diameter at an arbitrarily chosen conversion, viz. 90 %, increases
markedly with [Na +) (Figure 5.4). The minimum rNa oJ indicated in Figure 5.4
by the dotted vertical lines would represent polymerizations in the absence of
sodium carbonate. Higher values for [Na+] than the ones shown render latexes
with a poor shelf stability. The range of [Na + 1 in these experiments coTresponds
to a variation in ionic strength roughly between 0.1 mol.L·1 and 1 molLl, about
1 - 2 orders of magnitude larger than in the study of Goodwin et al.5,6) on the
emulsifier-free emulsion polymerization of styrene.
E 300 c: g " 250
200
150
100
50
O-r~~r-~-'-'~r-~-.-.-.
0.0 0.2 0.4 06 0.8 10
[Na'J/(moLL')
Figure 5.4. Partide diamerer dwat 90 % conversion (d~ V$. the overall [Na + j, ill two SDS concentrations Ct; -= 324 g.L·I (e); CE '" 8.1 g.L'1 (0).
At C~ -- 8-1 gJ •. 1 the observed S·fold increase in diameter corresponds to
a decrease in N by some 2 orders of magnitude. Realizing that in the standard
54 Chapter 5
recipe 2 the [Na 'j = 0.3 moLL· l (in analogy with recipe 1), it is obvioLls that
limited coagulation must playa dominant role in the partide formation process.
The reason for the presence of a moderate excess of pota.~sium carbonate in
recipe 1, ami for comparison maintained in recipe 2, is threefold:
- Performance and solubillty of dresinate 214 is strongly pH-dependent;14)
pH must be kept constant between 10 and 11 throughout the polymerization.
- In industrial practice the polymer content is usually high, between 30 and
60 % by weight. Inert electrolyte, are often added to reduce latex viscosity
and thus maintain a reasonable fluidity. IS)
- rn certain applications large sized spherical particles are desired, e.g. as
impact modifiers in materials exhibiting crazing and/or yielding (ABS, HIPS,
toughened PVC). Since the electrolyte induced coa,b'Ulation is confined to
microscale and does not result in reactor fouling, this phenomenon can be
used aovantageoLlsly in the onc-step preparation of large PB-particles.16•17)
The data from Figure SA can be presented alternatively by plotting the
particle nurnber (calculated from d9Q) on a log-log scale versus CB (Figure 5.5)
The .Iopes of these apparently linear (owing to the limited range of CE) curves
at a fixed (Na+)lhus represent the emulsifier exponent y toward N (Table 52).
Table 5.2. Variation of the emulsifier exponent y toward N with (Na +].
[Na + l/rnol.L-l
exponent y 0.1
0.5
0.15 0.2
0-8 1.4
0.3 2.1
0.4 2.2
0.5
2.4 0.6 2_6
The minimum value y ;;;; 0.5 is quite close to Y'h '" 0_60, obtained when
coagulation is fully neglected. Promoting coagulation by raising [Na -!oj clearly
increases the dependence of N on [El- Values for the emulsifier exponent as
high as :I have heen reported j n literature for other systems. 1a) The coagulative
nucleation model developed by the Sydney-groupll) predicts a range of exponen
tial values (0.4 :': Y,h :': 1.2), while raising the coagulation rate coefficient in the
nucleation model of Song and Poehleinl2) also leads to higher values for Y'h'
rolymerizations with Sodium Dodecylsulfate 55
~ .§. 100
e)
~~. z '" b 60
40
20
10
6,0
4,0
2,0
1.0
0-6
0.4
:/.~:
"~::l0 Figure 5.5. Pal1ic/e number N (calculated using d...) vs. SDS cOhcentration C~ as a function . ... oirlle overall fNa+ ] = 0.1 mol.
.~ L·1 (6); 0.15 molL' (.;: 0.2 moH.-1 (n); 0.3 moLL- (ej; 0.4 mol.L-1 ('\1); 0,5 molL! (A); 0.6 moiL! (0). AlTows indicate the maximum C E at that particular fNa+]. Data points ... at C~ = 32.4 g.C!, CE = 16.2
0 g.L·1
, and C. '" 8.1 g.L-1.
10 20 40 C,/{g,L i
)
Table 5.3. Influence of the overall fNa' ] on the p()iymerizaJion with CIS '" &1 g,£,1 SDS and Ill'" no mmoLL' (recipe 2).
[Na+] Hf.Rpol 1O-14.N/mL-1 .) ~ at 90 % p rnol.V1 g.mL-l.s-l DLS TEM nm
0_1 1.55 60 68 52 1.06 0.3 1.34 4,7 5.2 121 1.06 0.5 0.53 0.96 0.92 215 1,06
.) The particle numbers N determined with DLS are b:l,.Sed on d,.,. and those with TEM on d,,-
The effect of limited coagulation on conversion-time history (Figure 5.6) is
rather complicated: increasing INa +] gives higher initial rates in interval I, but
the steady state rate RpO' decreases significantly (Table 5.3). Similar results were
found with dresinate 214. This will be discussed somewhat further in chapter 6_
56 Chapter S
#. 100 c c v 0 80 #0 .~
/l <l) ;> /~1' <:: 0 0 60 0/ ..
/ .tJ../ o?v .tJ../ / .tJ../
40 ~7~ a/v
20 /J~v/ ...-:: .. ~v
~:;;,::o9 $-:::-0
.Q"'--.~
0 2 4 6 8 10 12 14 Time in hours
Figure 5.6. Conversion-time curveS jor polymerizations (recipe 2) with C" "" 8.1 g.L'] SDS, as a function of the overall fNa+ } = 0.1 moU,,-J
(\l); {N(1+ f = 0.3 mol.L·1 (0); INa+ I "' 0.5 molLl (n).
The average rate per particle for the~c and previous experiments with
dresinatc 214 (chapter 4), Can be calculated from the steady state RP<>l and the
final particle number N after cessation of coagulation. These values for ~l/N
are In fact upper limits for the individual experiments involved, since at lower
conversIons Rpot/N is smaller because of the smaller particle size (Figure 5.3).
Making a log-log plot of RpOt/N versus d90 (Figure 5.7), clearly shows that the
particle growth kinetics in intervalll is internally consistent, and depends solely
on partIcle size. A similar plot with varying dresinate 214 concentrations (Figure
4_6) gave an apparent linear relationship (owing to the limited data range); this
behaviour is seen to be encompassed by the generalized curve in Figure 5.7.
The graphs in Figure 4.6, Figure 5.3 and 5.7 are analogous to the well-known
Ugelstad·plots of log n vs. log , •. 19) Obviously RpoJN is equivalent to "_
Combining "-' '" P,v INk, and N '" V Iv (for definitions sec chapter 2), it is easily
seen that ,,' '" v2 0; db. At constant [I] (implying a constant p;), log (l<' can thus
be replaced by log d'l() or log d. A more rigorous treatment of particle growth
kinetics in intervals II and III will be given in chapter 7.
Polymemations with Sodium Dodecylsulfate 57
60 /' 00 <l0-rn 40 / , z '" I
8. 0
a; 20 / N 0
... 10 /
0
6.0 / 4.0
2_0 I' 0
1.0 40 60 100 200
doo/nm
5.4 Effect of the initiatOr concentration
Figure 5, Z Average /'ale per particlt: R~ vs. meQ11·volume diameter dv
aI 90 % CotWerSlon (d~. Final particle diameter was varied by changing the concentration of dresinaJe 214 (fl.). SDS (0) or the overall INa' J (\1).
The sodium persulfate (SPS) concentration was varied lOO-fold between
26.0 and 0,26 mmol.L"\ at two emulsifier concentrations, viz. C. = 32.4 g.L- l and
CE = 16.2 g.L'I. The INa'] = 0.3 mol.L'l and kept constant, so coagulation
phenomena may be considered a constant factor. The experimentally determined
initiator exponents are given in Table 5.4. The small effect of II] on Rp<>j is
rather typical of a Smith-Ewart case 1 system with n « 0.5 (see chapter 7)_
As limited coagulation is promoted (by lowering CE or raising [cation}), the
effect of [I] on N diminishes, since the partide number becomes predominantly
detennined by coagulation. This effect appears to be relatively small in the first
series of experiments from Table 5.4, since the exponent for N (x '" 0.46), is
quite dose to the theoretical one (x1h ", 0.40), when coagulation is ignored.
58 Cbapter 5
Table 5.4, Illiliafor e;r:.pom:lIts with regard to N and Rpol m different emulsifier concentrations CE and overall calioll concentratiom ICalion},
Recipe C. [cation] x in N 0: [I]' x' in Rp<>1 oc [Ir g,L'! mol.L'1
2 32.4 0.3 0.46 0.04 2 16.2 0.3 0.16 0.09 I 32.4 0.4 0.02 O.OS")
.J Taken from chapter 4.
The insensitivity of Rp<>' 10 Ihe lOO-fold variation of [I] is remarkable; its
effect On conversioll-time history is even more striking (Figure 5.8, with C" '"
32.4 g.L'1 SOS). Increasing [I], especially at [I] > 1.0 mmol.L'1, appears to be
slowing down the polymerization in interval I, eventhough the steady state
polymerization rate in interval II ~teadily keeps increasing with [I) (Table 5.5).
Results at CE = 16,2 g,L,l SDS arc similar.
Table 5.5. The periods of time required to reach 25 % (t2~ and 50 % conversion (t~, and RpOI as a fUlIction of Ilf at Ce '" 32.4 g.L l SDS.
[I]
mmol.L· '
0.26 0.52 1.30 2,60 6.50
13.0 26.0
t25
hours
2.55 2.50 2.60 2.90 :3.30 3.65 3.90
t.$(]
hours
4.40 4.35 4.40 4.75 5.20 5.60 5.85
1.61 1.64 1.75 1.69 1.79 1.82 1.92
Polymerizations with dresinate 214 (Figure 4.2) behave 'normally' in that the
duration of interval I increases with decreasing [I]. The discrepancy may be
caused by possible impurities in the emulsifier. Eventhough dresinale 214 has
been dis proportionated to eliminate abietic acid-type of derivatives, small
amounts are still present and can affect the kinetics, eF.pecially at lower [IJ.
Polymeri:l:ations with Sodium Dodecylsulfate
;;,100 c t:: 0 80 .~ (l)
> t;;; 0 60 0
40
20
0 0 2 4 6 8 10
Time in hours
59
Figure 5.B. Conversion-time CUlVes for polymerizations (Tecre 2) with CE = 32.4 g.L" and differenl /1/ .. 26.0 mmolLI (0); [IJ -- 0.26 mmoLL'/ (Il.). Initiator I is sodium persul!ate.
The absence of rtJarders in 50S makes the observed behaviour even more
surprising. On the one hand, increasing [II sufficiently, may possibly lead to
substantial termination of oligomeric radicals in the aqueous phase, thereby
reducing the radical entry rate into the particles. However, the initiator
concentrations are not excessively high, so this effect appears unlikely. On the
other hand, the behaviour in interval I may simply arise from differences in
initial particle size. This was observed previously, when promoting coagulation
by raising [Na+] at a fixed C. gave higher initial rates (figure 5.6).
This is further illustrated in Figure 5.9, showing particle number and
conversion-time curves for two polymerizations with a different [I] and IE]. hut
~th the same steady state Rpol and final N. The initial behaviour in interval I
was found markedly different, apparently as a result of a difference in colloidal
stability. The data in Table 5.5 suggest an analogous behaviour; increasing II]
above 1.0 mmol.L'1 clearly prolongs interval!. At [I] -- 13 mmol,L-l the particle
number was indeed found to decrease initially, whereas the polymerization
depicted in Figure 5.9 at [1] = 0.52 mmolLt shows no limited coagulation. So
colloidal stability is not only determined by the emulsifier concentration and
ionic strength, but also by the radical production rate (or alternatively (I]).
60
"* c 80 c 0 'iI) ill 60 :;. c 0
(.)
40
20
0 0 2 4 6 8 10 24
Time in hours
C,hapter 5
~ 5,0 ~
4,0
30
2,0
1.0
z ~-
(:)
Figure 5.9. Conversion (open symbols) (Ind panicle number N (full symbols) vs. rime curves for two polymerizations (recipe 2) with the same Rpol and /inal N: Ct "' 32.4 g.L-1 and [IJ '" 0.52 mmol.L'1
(triangles); Ce ;;= 16.2 g,Ll and II! '" 13.0 mmol.Lo/ (circles).
If the total polymer-waler interface increases faster than it adsorbs emulsifier
molecules, the surface charge density decreases and particles may become
colloidally unstable. A rapid reduction of surface charge density by incorporating
a high molecular chain into a small nucleus (I.e. rapid polymerization) was
theoretically deduced by Goodall et aL,20}
At a sufficiently high initiator concentration colloidal instability is apparently
induced. presumably because the emulsifier adsorption cannot keep up with the
rapid increase in particle surface as a result of the high particle nucleation rate.
Limited coagulation i~ observed experimentally, and consequently the reaction
slows down in interval L
This behaviour is also reflected in the apparent dependence of Non C". For
polymerizations with [I] = 13.0 mmol.L'l and [Na+] '" 03 mol.L,l a value of 2.1
wa~ found for the emulsifier exponent y (Table 5.1). However, decreasing [I] at
c;OnSlant [Na+] in the same range of [E] will give lower values for this exponent,
Polymerizations witb Sodium Dod~lsulfate 61
since the initiator exponent x at C. '" 32A gL-t and C. = 16.2 g.L-1 is different,
x '" 0.46 and x = 0.16, respectively (Table 5.4). Values for the exponent y
between 2.2 (for {I] = 26.0 mmoI.L·1) and 0.8 (for [I] = 026 mmol.V) were
calculated. Although these values for yare based on only two data points per [f],
the trend is considered significant.
S.S Comparison between dresinate 214 qnd sodium dodeqlsulfate
The two emulsifierS used in this investigation differ in almost every relevant
aspect: chemical structure, composition and purity, nature of the ionic group etc_.
Yet, this does not affect the reaction kinetics in interval II (and HI), as is
evident {J;-om Figure 5.7, which unambiguously shows that RpO/N is independent
of the choice of emulsifier. The prime function of the emulsifier in the present
system is thus the stabilization of latex particles formed in interval L
The emulsifier exponents for 80S and dresinate 214 were found to be 2.1
and 1.6, respectively (Table 5.1). However, in the experiments with dresinate 214
(section 4.3) the contribution of the emulsifier to the cation concentration had
not been taken into account; the overall [K+] thus varied with IE] between 0.3
and 0.4 moI.L-t• Additional experiments with a varying concentration of dresiml.te
214 at II fixed overall [K+] -- 0.3 moLe· were performed and compared with
similar experiments with SDS at [Na+] '" 03 mol.L-t (Figure 5.10).
The emulsifier exponents as deduced from the linear portions·) of the log N
vs. log CE curves are identical within experimental error, viz_ y '" 2.0 for
dresinate 214 and y = 2.1 for SDS. This indicates that the magnitude of limited
coagulation must be comparable, since a higher value for the emulsifier
exponent was earlier shown to correspond to a larger extent of coagulation.
0) It is generally recognized that over a wide range of Cs the log N vs. log CE curves are S-shaped for sparingly water-soluble monomers; this is beginning to manifest itself in Figure 5.10.
62
:..., 20 E ........ Z ~. 10 ;!
6.0
4.0
2.0
1.0
0.6
0.4
0.2
0.1
/' 8
II , I
II 1:1
•
0
4.0 6.0 10 20 40 60 C,/(g.L·')
Chapter 5
Figure 5.10. Particle number N versus emulsifier concentraJion C,Jor polymerizaliol'lS with SDS (0) and dresinate 214 (e), at a pemdfate concentration [ll '" 73 mmoLVl and an ovemll cation cOllcentmtion of 0.3 mol.L-i _
Table 5.6. Kinetic data for pOlymerizations with 16.2 g.L-1 SDS (recipe 2) and 32.4 g.L·1 dresinat(;! 214 (recipe 1), under othC!rwi.se comparable conditions. 0)
Emulsifier t b)
140 40 l<Y.Rpol !\, (run) 1O"1~.N 2(1
hours hou,s hours g.mL-1.s-J at 90 % mL-1
rjresinate 214 2.85 4.45 5.85 1.70 73 U SDS 4.75 6.75 8.20 1.68 74 2.1
oj Experimental conditions: (cation) = 0.3 woLL'! and [I) = 13 mrnol.L·1. b) t" Represents the time required to reach xx % conversion.
For a further analysis of the stabilizing potential of both emulsifiers the
polymerizations with 32.4 g.L-! dresinate 214 (ca. 95 mmoLLJ) and 16.2 g.C1
SDS (56 mmoi.L-1) were chosen, as the finaJ particle numbers were fortituously
found identical (Table 5-6). Again interval II kinetics are independent of the
type of emulsifier. The difference in the initial rates may be caused by
Polymerizations with Sodium DodecylsuU"ate
differen4;es in rate of emulsifier adsorption, or in the number of micelles initially
present. The conversion-time CUNe of the polymerization with 16.2 g.L·' SDS
was already shown in Figure 5.9.
The specific area per emulsifier molecule AE was determined by the method
of soap titration described by Maron et al}l) In short, the latex is titrated with
the corres(l<lnding emulsifier solution (including K+ /Na+ carbonate; [COJ:!-] "'
0.15 mol.el ), until the CMC is attained. Surface tension measurements with the
Wilhelmy plate method were used to determine the end point, which, in contrast
to Maron's findings,22) could be obtained accurately in this way. In fact, conduc
tance measurements as suggested for rosin soap solutions lZl are useless, because
of the high electrolyte 4;oncentration in the present system. Assuming that at the
end point the partides are completely covered with a monolayer of emulsifier
molecules, A" was calculated as:
6 MW (d:/d.}) P (E • CMC) p~ NAy
(5.3)
where P and E are the polymer and emulsifier content at the end point,
respectively, both in grams per unit volume of water, d. the mean-surface
diameter, d,. the mean-volume diameter (hoth obtained with TEM), Pr the
polymer density, and MW the molecular weight of the emulsifier. The main
components of dresinate 214 have the formula CI9H29C02K, giving MW = 340.
The CMC for dresinate 214 was found to be 1.2 g.L-1 (with (J;C03J = 0.15
moI.L"\ See se4;tion 4.3), and for SDS 0.17 gL\ ([Na~C03J ;;;; 0.15 moiL').
The value for A. of dresinate 214 (Table 5.7) is in excellent agreement with
the one of 43.7 lO"~o m2 determined by Maron et aL22) fo, dresinate 214 and
dresinate 731, the latter being the sodium salt analogue. The Sllrfa4;e coverage
of both latexes at 100 % conversion is readily calculated as (EjA.N,,/ .. d}N,
Since the emulsifier concentration [E] is far above the CMC, a corre4;tion for the
amount dissolved in the aqueous phase is marginaL This also makes a knowledge
of the full adsorption isotherms superfluous, so that the less cumbersome soap
titration method was considered sufficient for theSe systems.
64 Chapter S
Table 5.7. Characteristics of SDS, dresinate 214, and latexes prepared with 16.2 g.C I SDS and 324 g.L·! dresinate 2J4 (sec text).
Emulsifier lifo.A" Surface coverage Zeta-potential") m~ final latex in % mV
dresinate 214 45 68 - 80 sns 42 38 - 86
a) Determined at 25°C with a Malvern Zetasizer 3, with /Cd <:: 0.01 (sce section 5.3); dilution witb N~C03 solution (0.15 M).
1'he large differencc in surface coverage at tbe samc final N, clearly shows
that SDS is more effective in stabilizing PB-panicles than dresinate 214. This is
further substantiated by the dose identity of the zeta-potentials (I,") of both
latexes. The r potential represents the electrostati;;; potential at the 'slipping
plane' of the particle when it is in motion, and is readily obtained from electro
kinetic experiments (in this ;;;ase with laser Doppler spectroscopy of the particles
in an applied electric field). It is ;;;omparable in magnitude to the potential in
the socalled Stern plane (using a model for the double layer more sophisticated
than the one described in section 5.3), and is therefore a good indication of the
electrosta.bility of the particles. Apparently, the surface charge density per unit
amount of emulsifier ad~()rbed on (equally sized) particles is signifiGantly larger
for SDS. Although experimentally unambiguous, this effect is not readily
expl .. ined. Dresinate 214 contains ca. 10 % of non-surface active material. This
not only decreases the effective amount of surface active material, but it may
also interfere directly with the adsorption characteristics. The hydrophobic tail
of the main components of dresinate 214 consists of a rigid tricyclic ring system,
which definitely will affect the mode of adsorption, as compared with SDS with
its flexible linear taiL Altogether, the dresinate 214 emulsifier system is too
complicated to des;;;ribc its colloidal behaviour in the present context.
Polymerizations with Sodium DodecylSlIlfate 65
5.6 Condusions
Electrolyte induced limited coagulation is observed in polymerizations with
sodiulIl dodecylsulfate and dresinate 214, two emulsifiers differing widely in
chemical structure and composition. Sodium dodec:ylsulfate (SDS) was found a
more effective stabilizer than dresinate 214, because the surface charge density
per unit of amount emulsifier is appparently larger for SOS.
Coagulation is well established in polymerizations of polar monomers and
emulsifier-frel; systems; in cOlJ.trast butadiene is only sparingly water·soluble,
while [E] » CMC It is obvious that the experimental observations are not
specific for butadieue, suggesting that limited coagulation can be induced in any
emulsion polymerization system by electrolyte addition. Constancy of particle
number after the disappearance of emulsifier micelles seems more coincidental
than characteristic, strongly depending on the choice of reaction conditions.
The average rate per particle in interval If was found to depend solely on
particle size, and not on recipe parameters utilised to manipulate the final
particle diameter, such as the amount and type of emulsifier and thl; cation
concentration. No constancy of ii was found in the particle size range covered.
Promoting (;oagulation by raising [Na·jleads to a marked decrease in N,
while the exponent y in the empirical relationship N Q: [ll'.[EV increases.
Experimental values for y between 0.5 and 2.6 were found.
The dependen(;e of N on initiator concentration [f] is also strongly
determined by the colloidal prope(ties of the system. Increasing coagulation
reduces the e1Cponent x toward zero, since the final partide number becomes
independent of the nucleation rate of (colloidally unstable) primary particles.
Raising [Ij at a constant ionic strength increases the exponent y over a
narrow range of C~. Apparently, adsorption of emulsifier on newly formed
particles lIlay become rate-determining, resulting in a limited colloidal stability.
Kinetic relationships between particle number. [XJ and [E] appear 10 be
extremely complicated in systems with a limited colloidal stability. Nevertheless,
electrolyte addition is common practice in industry (e.g. for freezing point
depression, or to decrease latex viscosity), so that a better understanding of its
effects on reaction kinetics and mechanisms may prove extremely useful.
66 References
1 CP. Roe, Ind. Eng Chem. 60, 20 (1968) 2 WV Smith, R.H. Ewart, 1. Chem. Phys. 16, 592 (1948) 3 K. Wendler, L Pielert, M, Fedlke, Plaste Kautsch 30. 438 (1983) 4 R. Buscall, R.B. Qttewill, 'The Stability of Polymer LaJices'~ in: "Polymer
Colloids", eoiled by R. Buscall, T. Comer, J.F. Stagernan, Elsevier Applied Science Publishers Ltd., Barking 1985, p. 141· 217
5 J.W. Goodwin. J. Hearn, c.c. Ho, R.H. Qttewill, Brit. Polym, J., 347 (1973) 6 J,W. Goodwin, R.H. OUewill, R. Peiton, G. Vianello, D.E. Yates, Brit,
Polym. j. 10, 173 (1978) 7 A.S. Dunn. L.CH. Chong, Brit. Polym. 1. 2, 49 (1970) 8 R.M. FilCh, C.H. Tsai, "Polymer Colloids", edited by RM. FilCh, Plenum,
New York 1971, p. 73 9 V.I. Yeliseyeva, ''Polymerization of Polar Monomers", in: "Emulsion
Polymerization", edited by I. Piirma, Academic Press, New York 1982 10 F.K. Hansen, J. Ugelstad,1. Poly. Sci., Polym, Chem. Ed. 16, 1953 (1978) 11 P.J. Feeney, D.H. Napper, R.G. Gilbert, Macromolecules 17, 2520 (1984) 12 Z. Song, G.W. Poehlein, 1. Mw:romoL Sci. " Chern. A25, 403, 1587 (1988) 13 CD. Reed, J.J. McKetta, 1. Chem, Ens. Data 4, 294 (1959) 14 D.C. Blackley, "Emulsion Po/ymerisation • Theory and Practice", Applied
Publishers Ltd., London 1975, p. 291 15 Ger. 26 45 082 (1976), Chemische Werke Hills AG, Inv.: H. Schleuter;
Chern Abslr. 89, 25805s (1978) 16 Czech. CS 200 436 (1978), Inv.: D. Konecny, S. Bittner, S. Mysik, J. Svarc,
1. Trneny; Chern. Abstr. 99, 71329u (1983) 17 JP 59 22 904 (19S2), Japan Synthetic RUbber Co. Ud.; Chern. AbSlr. 101,
39665z (1984) 18 RM. filCh, Brit. Polym. J. S, 467 (1973) 19 J. Ugelstad, P.C MOrk. 1.0. Aasen,1, Polym. Sci.: Pan A-IS, 2281 (1967) 20 AR. Goodall, M.e. Wilkinson, J. Hearn, "Characterization of Pwticies
during Growth in Emulsifier·free Emulsion Polymerization of Styrene", in: "Polymer Colloids II", edited by RM. Fitch, Plenum Press, New York 1980, p. 629 • 650
21 S,H. Maron, M.E. Elder, I.N. U1cvilCh, 1. Colloid Sci. 9, 89 (1954) 22 S.H. Maron, M.E. Elder, C. Moore, J. Colloid Sci. 9, 104 (1954)
Physico-Chemical Asp«ts 67
Chapter 6 Physito·Chemical Aspects
SUMMARY: The stirring speed n affects the emulsion polymerization of butadiene, (1) by reducing the effective emulsifier concentration [Elect available for particle nucleation and stabilization at high n, and (2) by limiting diffusion of monomer to the latex particles at low n, The large density difference between butadiene and water promotes the breaking up of droplets at high n, while the same condition constitutes a large driving force for (partial) phase separation at low n. Increasing the monomer/water ratio at constant lEI decreases [E].(I'> and thus the final partide number. At monomer volume fractions > 0.6 mixed emulsions are likely to be fonned initially, reducing [EI.IT even further. In the presence of mixed emulsions, polymerization in the monomer phase may no longer be neglected, giving rise to a complex kinetic behaviour.
6.1 Introduction
In previous chapters the role of various chemical and colloidal parameters
in the emulsion polymerization of butadiene was discussed. However, systems
using monomers at concentrations above their saturation water solubility are
necessarily heterogeneous in appearance, due to the presence of a separate
monomer phase in intervals I and II. The physico-chemical properties of the
monomer emulsion may markedly affect the course of the polymeri~tion.
1111; effects of stirring on emulsion polyrnerhmtion is a sornewhat neglected
aspect in this line of research; only a few studies have been reported. t ,z,3)
The main conclusions reached in these studies include: (1) the effective
68 Chapter 6
emulsifier concentration [EJd[ in the aqueous phase may be reduced by
ad~()rption of emulsifier molecules onto the monomer droplets, especially with
[E] '" CMC; (2) at low stirring ~peeds transport of monomer from the droplets
to the aqueous pha~c may become rate-determining due to (partial) phase
separation; (3) at high stirring speeds particles may coagulate and coalesce.
The influence of stirring in the present study, together with the effect of the
monomer/water ratio, will be discussed in the following. It will be shown that
the atypical phy~ical properties of butadiene, being a liquified gas under the
prevailing experimental conditions, accentuates the significance of the physical
state of the polymerization system.
6.2 Agitation
Polymerizations were performed in a 2.3 litre reactor fitted with four baffle
plates located at 90" intervals and a tweJve flat-bladed turbine impeller (see
Figure 3.2). In two series of experiments, one with 32.4 g.L·1 dresinate 214
(recipe 1) and another with 16.2 g.L-1 SDS (recipe 2), the stirring speed n was
varied and the final particle number N determined (Figure 6.1). In all cases the
system was preemulsified by stirring a few minutes at 400 rpm, before adjusting
n to the desired leveL In both reaction systems, at sufficiently high n the particle
numoer was found c(m~tant within experimental error. while a discontinuous
increase in N became apparent when going to lower values for n. Although
reproducibility of ah initio polymerizations is not always satisfactory. especially
at low n, thc change in N is considered significant. As pointed out earlier, the
effective cmul~ifier concentration IE)cff available for particle nucleation and
stabilization is influenced by adsorption of emulsifier onto the monomer
droplets. I\t high n the droplets will be mOre finely dispersed and IE) > [EloIT
Going to lower n, agitation ultimately becomes insufficient to counterbalance
droplet coalescence, so that [E10fl' and thus N, increase. Since the change in N
is discontinuous, the phenomenon must be related to the flow conditions.
Physico-Chemical Aspects 69
~ 2.4 ~ "-,
~ 2.0 b T-
1.6
L2
0.8
0 200 400 600 800 1000 stirring speed in rpm
Figure 6.1. Final pUTtide number N vs. stirring speed, for polymerizations wah 324 g.L-J dresinate 214 (0) and 16.2 g.L-l SDS (.).
The impeller Reynolds number Re, defined as the ratio of ine(tial forces
versus viscous forces, can be used as an indicator of the overall conditions of
tlow. In the case of a mechanically agitated system, Re = D2npfJj, where D is
the impeller diameter, and p and Jl the density and dynamic viscosity of the
liquid, respectively. At Re > 104 flow is usually highly turbulent, which in the
case of a liquid-liquid system of two poorly miScible liquids wili result in a
homogeneously distributed dispersion. If the two liquids have different densities
and viscosities, the volume ratio of the two phases also needs to be considered.
The average density and viscosity can be calculated a<:cording to:4,»
p <PcPc + 'Polin (6.1)
JlG [ 1.5JlI>IP" ] Jl=- 1+~~-<Pc I'D + /Jc
(62)
whe(e If' is the volume fraction, while the subscripts c and [) denote the
continuous and the dispersed phase, respectively.
For the polymerizations depicted in Figure 6.1 with 'PD '" 0.43, n :: 230 rpm
was calculated to provide sufficient turbulency. Agreement with measured values
for n above which N is constant is reasonably good, taking into account the
unsatisfactory reproducibility of these ab initio polymerizations at low n.
70 Chapter (0
U~ing the impeller Re to characterize the flow field in an emulsion
polymerization system is of course of limited value, since the presence of the
emulsifier and its ability to suppress coalescence of monomer droplet5 is
ignored. However, as polymerization proceeds emulsifier molecules become
adsorbed on the latex particles and less is available for stabili~ing the monomer
droplets. At the end of interval I the monomer emulsion thus tends to become
unstable and mechanical agitation is essential to keep the mOnomer sufficiently
dispersed. Inadequate stirring at that stage inevitably leads to a reduction of the
total surface area of the monomer emulsion; even creaming-up and (partial)
de mixing might occur. This will inevitably affect the polymerization rate as
diffusion of monomer from the droplets into the aqueous phase become~ rate
determining and the monomer concentration within the latex particles decreases
below its saturation value. This is evident from Figure 6.2, showing conversion
time curves for polymerization:; with 32.4 g.L·) dresinate 214 at 100 rpm and 400
rpm. The curve at n = 100 rpm is based On a limited number of data points, as
representative samples could not be taken due to the inhomogeneity of the
system. Each data point at n .. 100 rpm stems from a separate polymerization;
the <;Qnversion was determined by raising n to 400 rpm, followed by sampling.
*' 100 .S c .Q 80 ~ ill > c 8 60
40
o 20 40 60 80 Time in hours
Figure 6.2. Conversion-time curves for polymerizcUiow,' with 32.4 g./...J dresinare 214 (recipe 1), at a stirring speed n '" 400 rpm (unbroken line; dtUa points omitted), and 11 '" 100 rpm (0) .
Physico-Chen:lical Asp~ls 71
For polymeriMtions where n is raised from 100 to 400 rpm, the development
of the particle number aftt:r the increase in n is shown in Figure 6.3. In all cases
N decreases to a level corresponding to a polymerization with n = 400 rpm from
the start of the reaction. Redispersion of the monomer by raising n creates new
interface, whereupon emulsifier molecules adsorb. Thus emulsifier desorhs from
the Iate1< particles, introducing colloidal instability and a decrease in N.
~ 2.5
E --:z; 2.0 " b
1.5
1.0
0.5
0 20 40 60 80 100
Conversion in %
Figure 6.3. Partide number N vs. ccmversion for polymerizations (recipe 1) where the stirring speed is raised from 100 to 400 rpm
at differem conversions.' 18 % (0), 37 % (t:.) and 50 % ('\1).
The pressure as sensed with a pressure transducer can serve as a qualitative
indicator of the state of mixing, In a properly mixed system the pressure is equal
to the saturation pressure of butadiene vapour, and will remain virtually constant
until the beginning of interval III (Figure 6.4a). At n = 100 rpm a different
behaviour is observed; pressure drops below the butadiene saturation pressure
well before the end of interval II, remains more Or less constant thereafter, and
drops off sharply at the end of the polymerization (Figure 6.4c). Upon raising
n from 100 to 400 rpm somewhere in 'interval II', the pressure rises immediately
to saturation pressure and behaves normally after that (Figure 6.4b).
When stirring is insufficient, mass and heat transfer into the gas phase is not
optimal, while at the same time butadiene vapou( is continuously condensing
72 Chapter 6
against the (non jacketed) top of the reactor. Consequently, the temperature in
the ga~ phase will decrease somewhat, and pressure cannot be maintained at
saturation level.
'~I : ITTT-rl--,]"" ,! lI8--r
I'" I ' . 1 I
I I I I I ' L, ______ '_'- -:1 .-. --f--- -1-
r .1
~--~---!t,--i,··
, I-'I-r-- -T1T~- ---- . \-I-I-~-- ---,-- ---Ir-i
-- -: ++1-/ ---- -- -.- i Vi-if --- --~j- ---- -1 ~I Il-r-rJ" ~ + ---- ..
I'
n,,1 i JJ ___ L_i_i
, -I· '_I! LI __ :
: 1 :
i ' :_.i : I '
I : I ,I, 1
:- )aJILII~:-I·I- ~ i I : i I . I i .. L_._~.LL_-,_ .... I .. I
- r--' ,
- I --+-+-+-++---'f-----+--- -
1-+-\---"; - -----." --- -+-1- • -- -- --f- --+--
'-+-I-~-+"+-I-+---1--I--l--- -- --~
-+-+-++-S-- -- (b)- '-+-----_ L ___ J_I_. _________ -.1
.. -----f- ,,--, -.r 1--- - .'-~j
--, ,···f--f---I---+-t-+-f-il+-·· .. +-+-+-1-\-+-+- --I;-F.<-----Temperature- , ... I-'-----H'-t-+-+-+-f--- .. -- -::::p i '
Figure 6.4. Recordings of tellveruture and pressure (read ham ril/hi to left) for polymerizations with 32.4 g-C drl;sinute 214, at a stirring speed n = 400 rpm (a); n raisedfrom 100 rpm to 4()(J rpm, indicated by the arrows (b); n = 100 rpm (c:).
Physico-Chemical Asp~1s 73
For dispersions of two poorly miscible pure liquids an empirical correlation
for the total interfacial area per unit volume of dispersed phase AD> was
obtained by Rodger et al.:6)
(6.3)
where K and K are constants, 0 the impeller diameter, D, the reactor diameter,
t the settling time of the dispersion, to the reference settling time, "I the
interfacial tension, t:.p the density difference, II the kinematic viscosity (= Il/p),
and 4> a scale-up function. This relation is in accordance with the model for
agitated dispersions proposed by Shinnar and Church,7) based On Kolmogoroffs
theory of local isotropy in turbulent flow fields.
A correlation for the minimum stirring speed nmin for complete dispersion
of two virtually immiscible liquids has been given by Skelland et aLa)
(6.4)
where K' and K' are constants, and g the gravitational constant
Stdcktly speaking these correlations are only valid for emulsions of pure
liquids, and cannot be used to describe an emulsifier-stabili2:ed monomer
emulsion. Yet, they do imply that t:.p, and to a much lesser extent II, also need
to be conSidered with emulsifier-stabilized emulsions, since both parameters are
macroscopic properties. The emulsifier will exert its effect on A" mainly by
reducing the interfacial tension 1 and increasing the settling time t. The unique
physical properties of liquid butadiene (b.p. '" - 4,4°C) tend to increase nmin and
A" (i.e. redul;e the average droplet diameter), relative to other monomers (e.g.
styrene; b.p. '" 145°C) under otherwise identical conditions (Table 6_1).
It is evident that a large relative density difference Ap/pc promotes the
breaking up of dwplets in a turbulent flow field. Furthermore, in the presence
of emulsifier coalescence is greatly suppressed, consequently a rather small
average diameter of the monomer droplets is expected in the present s)'5tem.
74 Chapter 6
Table 6.1. Physical properties of butadiene, styrene and water at 6:rC.
pure liquid 104.1' 107.1' P t:.p/p,')
kg.m·l.s·1 m~.s·1 kg.m·3
butadiene 0.939) 1.65 565 0.42
~tyrene 4.58 5.28 868 0.12 water 4.53 4.61 982
0) For monomer emulsions in water.
The average droplet diameter of a monomer emulsion d" can take values
between 5.10.5 m and 1.10"" m, depending on the method of emulsification. lO)
Based on the previous discussion d" of a butadiene emulsion in a turbulent field
is expected to be found at the lower end of this scale. This was experimentally
verified by an Os04-staining of a butadiene emulsion prepared at 400 rpm with
16.2 g.L·1 SDS. The procedure is a modified version of the technique described
by Yang. lI) In short, a sample of the initial emulsion was taken with a 10-4 dm3
OLe 8ampling valve, and immediately « 5 sec) contacted with OS04 by pushing
a 2 % Os04-wlution from one high-pressure-pTOof syringe to another via the
GLC sample loop. The solution turned btack within seconds. The resulting
stained particles are ready for electron microscopy analysis with SEM and TEM
(Figure 6.5). Particle counting gave an average diameter do .. 8.10-7 m. Although
the uptake of OS04 will certainly affect density, the order of magnitude is
considered typical of the initial droplet diameter in all recipes, as [El » CMC,
while the breaking up of droplets is mainly determined by the macroscopic
property 6.p/ p,;;.
Using A" = 42 W-20 m2 for the specific area per molecule SDS (section 5.5),
the effective emulsifier concentration in the aqueous phase, for the recipe with
16.2 g.L·1 SDS varies with dM as shown in Figure 6.6. With d .. '" 1O.{) m, it follows
that ca. 30 % SOS is initially adsorbed onto the droplets, despite the fact that
[E) " CMC (= 0.17 g.L·1, section 5.2). In contrast, Nomura et al?) concluded
for the emulsion polymerization of styrene with SDS as emulsifier, that
emulsifier adsorption on monomer droplets and the effect of slirring thereupon
were both insignificant for recipes with [E] ::: 12 CMC.
Physico-Chemical ASpects 75
Figure 6.5. Transmission electron microsraPh (left) and scanning electron micrograph (right) of an OsO.stained butadiene emulsion, prepared at a stirring speed n = 400 rpm, in the presence of 16.2 g.L,1 SDS (recipe 2).
* 100 ,S;
60
40
20
0.1 0.3 0.6 1.0 3.0 6.0 10 30 60 100
10'.d./m
Figure 6.6. .Effective ~mulsifier concen~ration [E11
V.I'. droplet diameter d" for a butadiene emulsIOn wIth 16.2 g.L 1 SDS ({£ = 56 J(j!J moLLl) and a mQnomer/waler rotio MjW '" 3/1 g.g.l (recipe 2), with [E] <If calculated
according to (I - 6(MjW)/«E] . CMC) A~ NAv d" p",} 100 0/0.
76 Chapter 6
It should be noted that such finely dispersed butadiene emulsions are
unavoidable, as the relatively high stirring speeds are necessary to obtain
homogeneous emulsions, allowing reproducible sampling. Yet, the effect on
reaction kinetics in the present system is propably small, as partide formation
is coagulative in nature, and stops at 40 • 45 % conversion, when most of the
emulsifier has already become adsorbed onto the particle ~urface.
If the total surface area of the monomer droplets is sufficiently large,
polymerization in the droplets might no longer be negligible. Smith and Ewart1Z)
conveniently ignored radical entry into monomer droplets, as emulsification
normally gives values for d .. in the size range of 2 • 5 10.0 m; droplets of this
size cannot compete effectively with the smaller and more numerous micelles in
capturing radicals. A similar argument can be made for homogeneous nucleation
mechanisms. However, this argument might no longer hold in the present
system. Yet, the fraction of particles> 5.10.7 m (supposedly polymerized
droplets) normally found in the particle size distribution of the final late):; is
insignificant as compared with that of the true latex particles.
Several reasons can be given for this negligible droplet polymeri:;o:ation,
despite the large overall surface area of the initial monomer emulsion:
- The value for d" is the result of a dynamic equilibrium between the
breaking up of droplets by shear and viscous forces, and droplet coalescence.
The latter process is mainly determined by diffusional thinning of the films
surrounding adhering droplets. The high stability of emulsions in the presence
of surface active agents mainly results from an increased stahility of thi~
protective film. The experimentally determined droplet diameter is that of the
initial emulsion, when [E] » CMC, As polymerization proceeds. [El drops below
the CMC and A" must decrease due to a higher rate of droplet coalescence. The
number of droplets (and thus A,) further decreases quickly by diffusion of
monomer from the droplets to the latex particles .
• As droplets continuously coalesce there is a constant exchange of material
between them; the polymerization kinetics in the droplets i5 essentially that of
a bulk process, which is considerably slower than tbe corresponding emulsion
Physico-Chemical Aspeds 77
process. Only when the latter is e:dremely slow (usually at high monomer to
water ratios or at low N), polymerization in the droplets becomes competitive.
This shows up in the particle size distribution (PSD) of the final latex, which
begins to show an extreme positive skewness, with particles sometimes as large
as 10~ m; these particles evidently ~tem from polymerized droplets. An
illustrative example is shown in Figure 6.7, giving PSD's of final latexes prepared
with 8.1 g.L"! dresinate 214 (recipe 1) at different [K+] (Table 6.2).
Table 6.2. D(1/a on polymerization kinetics and pMicle size distributions for experiments with 8.1 g.L-l dresinate 214 (recipe 1) at different overall [K" I-
[K+]
mol.L-l
0.325
0.35 0.375
0.40
0) fCd..)
rJZ 100 ~
I: o 80 ~ ,g i; '5 60 Q} ;,.
~ ~ 40 E ::. o
20
time conv. 1W.Rpal d. at 90 % f(d.) > P
hrs % g.mi..l.S·1 nm 5.10-7 m 0)
25 96 6.9 181 l.O4 30 90 4.3 236 L05 72 94 1.4 358 1.14 14
120 93 0.88 375 1-49 35
'" the fraction of particles with diameter dy-
200 400 600 800 1 000
dJnm
Figure 6. 7. Cumu/alive particle size dislribution.s for polymerizations wilh 8.1 g.L-1 dresinate 214 at different overalJ {K+ J = 0040 moLL-l (0),- /K· J =
o.J? moLL'/ (~); {K' J '" 0.35 mol.L-1 ('V);{K+ J '" O.3~ moLL-1 (e). MjW =
3/7 g·t1 (recipe J).
78 Chapter 6
Similar results were reported for styrene emulsions prepared with ultra
~onification, with d" '" 10-6 m.lO) After polymerization, the weight ratio of large
partides/small particles in the final latex wa~ found negligible, despite the fact
that initially 25 - 30 % of the emulsifier was adsorbed on the monomer droplets
(estimation of the present author using ~ -- 45 10-:11) m2 for Aerosol MA13)).
Under turbulent flow conditions (Re > W'), butadiene emulsions with a
~mal1 d" can be prepared relatively easily, even with customary equipment, as
a result of the unique physical properties of this monorneL Paradoxically, at Re
.:: lit the large (>.p also promotes the settling out of the two phases, especially
in interval II when the droplet surface is slowly being depleted of emulsifier.
The dtlration of interval II for butadiene polymerizations i~ of the order of
several hours; besides a large driving force for phase separation, the opportunity
to do so is also overwhelming. Once started, the process of creaming and
demixing is irreversible <It constant n, ultimately leading to a diffusion·controlled
polymerization. This is in agreement with experiment<ll results (Figure 6.2).
An illustrative example of ineffective mixing found in literature, \S the
radiation-induced emulsion polymerization of butadiene reported by lshigure et
a1.. 14) Gla~s ampoules filled with butadiene emulsions were placed on a rotating
wheeL') In view of the foregoing discussion, this experimental set-up appears
inadequate for dispersing hutadiene properly. Indeed, ,a pronounced effect of the
agitation conditions was found when stirring was improved by using a magnetic
stirring bar; in ~ontrast similar polymerizations with styrene were daimed to
behave normally using the rotating wheel. This is consistent with our ob5ervation
that at Re < 104 phase separ<ltion is relatively facile for butadiene emulsions,
while emulsions of styrene under identical conditions can remain adequately
dispersed. Styrene, being more reactive than butadiene, will have passed through
the critical stages in interval II hefore extensive demixing could have occurred.
Simila'r comments apply to results on butadiene emulsion polymerizations
reported by Morton et a I.. I~J The socalled bottle polymerization technique was
.J The degree of filling of the ampoules and the gravitatiolJ<lI orientation ()f the wheel (believed to be vertical) were not specified.
Physico-Chemical Aspect<; 79
used, where bottles were rotated vertically Md-over-end at 35 rpm. The initial
degree of filling of the bottles was extremely high, about 90 %.'6) Experimental
results thus obtained, together with data fwm Wendler et alY) (using a 2 litre
reactor with n '" 360 rpm) and the present investigation, are shown in Figure 6.8
as a log-log plot of R.,.,JN vs. ~. The values for ~JN from Morton et al.!')
are significantly lower than those in the other studies where stirring was
adequate, and can only be partially attributed to the difference in temperature
(only 2'e with the present study) and the particle sizing method (soap titration
in ref. IS». Again, ineffective mixing appears responsible for this behaviour. It
may be noted that the value for kp reported by Morton et al. (100 L.mol·1.s·1 at
62°C),t8) is significantly smaller than our estimate, kp ::: 180 L.morl.s·1 (see
chapter 7). This is to be expected with a diffusion·contwUed polymerization; a
lower value for Rpol inevitably results in a lower value for Js,.
~ 100 '(h
rn ::::::: 60 2:
a:.t 40
b 20
10
6.0
4.0
2.0
1,0
t;> <;>
I • '" I
" 40 60 100 200
Figure 6.8. Average rate per particle R~N vs. particle diameter at 90 % t;onw:rsion (dgo), for poiymerizatiollS reported in this investigation at 62"C (0), from Wendler et al. l7
) at 7fJ'C (<:7), and Morton et aL IS) at 6()'C (/:.). Data of first two references based on volumeaverage diamelers d, and of ref- m on swface-average diamelers d •. Full circles indicate experiments with 32.4 g.L"l SDS at different monomer/ water ratios MjW (see section 6.3).
400
80 Chapter'
6.3 Monomer to water ratio
Commercial emulsion polymerization processes in batch-operation are for
obvious economical Tea~ons performed at the highest polymer content feasible.
For industrial butadiene emulsion polymeri:/;ations monomer to water weight
ratios (M/W) up to 1.5 are employed.19) As is obvious from the previous
di~Cl,lssion, the agitation conditions at such high M/W must be considered for an
optimal performance of the process. Table 6.3 shows the final particle number
as a function of the stirring speed n at M/W '" 1 with 32.4 g.L·! dresinate 214.
Emulsions appcar adequately stabilized by turbulence at n :> 400 rpm; with 'PD
= 0.64, an impeller Re of 10" is Obtained with n = 450 rpm.
Table 6.3. Effect of the stining speed n on patticie number N for polymeri4alion.t at M;lV '" J with dresinaJe 214 (recipe 1).
n in rpm lO,15.N/mL·1
200 0.38
300 0.42
400 0.22
500 0.21
750 0.20
Going to yet higher M/W it was realized that a turbine impeller might no
longer provide sufficient agitation. as it is typically designed fO{ low-viscosiry
systems (I' ::: 50 kg.m-ls-I). This was indeed experienced at M/W = J.5. where
power consumption increoe;ed drastically and temperature control was poor. For
these polymerizations a low pitch helical ribbon agitator woe; chosen at n = 300
rpm. The ribbon agitator and the turbine impeller at M/W ::: 1 were found
equivalent in performance with regard to conversion-time history and final N.
The monomer to water weight ratio M/W was varied between 0.25 and 1.5,
with both drcsinate 214 and sodium dodecylsulfate (SDS) as emulsifiers. All
other recipe ingredients were kept constant on water basis. Conversion"time
curves for polymerizations with SDS at val)'ing M/W are given in Figure 6.9,
and those of polymerizations with dresinate 214 at M/W '" l.5 in Figure 6.10.
The dependence of final particle number on M/W for polymerizations with both
emulsifiers is shown in Figure 6.11.
Physico-Chemical Aspects
"it 100
,,; c; -~ 80 ffi ~ 8 60
40
20
a 2 4 8 10 12 14 Time in hours
Figure 6.9. Conversion-time curves for polymerizations with 32.4 gL-1
SDS (recipe 2), aJ different mDnomer/water rmios MjW -- 0.25 (6); MjW == 3/7 (e); MjW '" 2/3 ('>1); MjW -- 1 (.t.); MjW = 1.5 (0).
<fi ,,; c 60 0 '~
~ '" 0
U
40
o 10 20 30 Time in hours
81
Figure 6.10. Conversion-lime CUIVE!S jor polymerizations with a monomer/ water ratio M/W -- 1.5, at different dresinate 214 concenlratiDItJ: CIS '" 61.6 g.L·1 (i:l.); CE ~ 51.8 g.L·' (0); CE '" 421 g.CI ('>1); C, '" 32.4 g.L" (0),
82
~ 20
.§. z 10 " 0 6.0
4.0
20
1.0
0.6
0.4
0.2
0.1
0.06
0.04 0.2 OA 0.6 1.0 2.0
M/W in (g.g")
Chapter 6
Figure 6.11. Particle number N vs. monomer/WalerratioMjw,for polymerizations with 32.4 g.L·1
drosinate 214 (0) and 32.4 gL1
SDS (L). Particle sizing by DLS (open symbols) and rEM (full symbols).
It should be noted that the volume fraction 'P of butadiene is considerably
larger than the (:oHcsponding weight fraction. For example, for M/W = 1, thc
initial 'I' = 0.64, and for M/W = 1.5 even 'P = 0,73, both at 62"C. At such high
volume fractions Ihc type of emulsion may differ significantly from the normally
encountered oil·in-water type (o/w). Rodger et al.'l) noted in their investigation
with pure liquids") that many o/w emulsions (:ould be inverted to the water-in
oil type (wio) by increasing the energy input, this being more facile with larger
value5 for I':.P/h. In fact, ~y$tems with ~p/Pc values of the ordcr of 0.6 could not
be m<l.dc to produce o/w-type emulsions without changing the position of the
impeller. For butadiene Ap/Pc = 0.42 (Table 6.1). However, the use of anionic,
water-soluble emulsifiers will favour emulsions witb a continuous aqueous phase.
Although a thowugh knowledge of the hehaviour of butadiene emulsions
scems crucial, a microscopic study is virtually impossible. due to toe low boiling
~) Volume fraction of the organic phase believed to be 0.5.
PhrsjCll-£hemical Aspects 83
point of butadiene (- 4A'C). In order to get some insight in the emulsification
behaviour at high M/W, styrene emulsiorn prepared with dresinate 214 at
different phase ratios were studied instead (Figure 6.12). This experiment
showed that the (handshaken) emulsions get coarser with increasing M/W, while
at styrene volume fractiorn ::: 0.60 mixed w/o/w emulsions were being formed.
Obviously, increasing 'P at a cornlant emulsifier concentration, must lead to
a decrease in [El.fl' and thus in final N. At M/W .:: 1 (i.e. '" :: 0.63), when mixed
emulsions are formed, [El.fl and N will decrease even further. This is clearly
observable in the experiments with SDS as emulSifier (Figure 6.11).
The type of emulsion may also affect the course of polymerization. As
interval II ends at 60 % conversion (see chapter 7 and Table 6.4), it is obvious
that many conversion-time curves at high M/W have two kinetic regimes in
interval II. In the presence of a mixed emulsion, polymeri~ation in the monomer
phase may no longer be negligible. As polymerization proceeds, the monomer
volume fraction 'P decreases, so that a trall$ition fwm a mixed emulsion to a
normal o/w emulsion is likely to occur, as the latter is preferred at lower ",.
This change in type of emulsion, which is expected to be abrupt, will reduce
radical entxy into the monomer phase significantly. Based 00 this concept,
polymerization before the rate inflection occurs simultaneously in the latex
particles and monomer droplets, while after the inflection the latter locus is
strongly reduced Or eliminated. Unfortunately, this hypothesis is extremely hard
to test, as butadiene emulsions cannot be studied in a direct way.
However, the kinetics of the experiments with 32.4 g.L-t SDS at different
M/W (Figure 6.9) provide some useful indicatiOIl$ (Table 6.4). The average rate
pel" particle R,..,,/N for these polymerizations is indicated in Figure 6.8 by the
full circles. The polymerization with M/W '" 1 is of special interest, as it clearly
shows two kinetic regimes. The rate Rpol used to calculate ~t!N is the One
after the inflection point. As can be Seen in Figure 6.8, the experiment with
M/W == 1 (having dw '" 90 nm) fits in perfectly with the rest; the second regime
appears to be that of a true emulsion polymerization (for the first regime
~t!N '" 5.6 1~1 g.S·I). On the other hand, polymerization in the first regime
is indeed faster than anticipated on the basis of the experiments at lower M/W.
84 Chapter 6
Figure 6.12. Optical micrographs of styrene emulsions (handshaken) in the presence of 32.4 g.L-1 dresinate 214, at different styrene volume fractions 'P = 0.20 (a); 'P = 0.30 (b); 'P = 0.40 (c); 'P = 0.50 (d); 'P = 0.60 (e-f);
'P = 0.70 (g); and 'P = 0.80 (h). Light areas represent styrene domains.
Physico-Chemical Aspects 85
Table 6.4. Kinetic data of po/ymerization.s with 32.4 gLI SDS (recipe 2) and different monomer/water ratios Mjw.
M/W 1!f·F),.,1 ~at90% p x' a) g-g'! g.mL-I.s-1 nm
0.25 1.1 0_593 0,43 1.8 44 1.06 0.598 0.67 2.6 60 1.07 0.588 1.00 1.5b)
3.4 90 1.06 0.607 1.50 2.1 197 1.30
0) cOnversion where interval III begins (see section 7.3). b) Rate before the inflection point.
Beside the latex particles, an additional locus for polymeri2:ation, VlZ. the
monomer pliase, must be kinetically significant in the~e systems.
The number of large off-sized particles found in the latex with M/W .. 1.5
was significant; the PSD showed a profound positive skewness, with particles as
Large as 6.10'7 m. This was not observed in the final PSD's at M/W < 1.
The polymerizations at M/W .. 1.5 with 61.6 g.L-! dresinate 214 (d" = 174
run at 90 % conversion) and 32.4 g.L-! 5DS both lack a clear inflection point;
both final particle si.res are comparable. Decreasing the concenlration of
dresinate 214 obviously leads to smaller particle numbers, while at the same
time a negative inflection in the conversion-time curves hecomes apparent
(Figure 6.10). A decrease in interval II rate was observed earlier in an
experiment with M/W '" 3/7, where the particle number was reduced substan
tially by electrolyte addition (Figure 5.6, with 8,1 g.L-! SDS and [K+j '" 0.5 M).
Polymerization in the monomer phase again appears competitive in the initial
stage of polymerization, this time caused by the extremely low particle number_
The leveling off at d90 > 200 nm of RpoJN in Figure 6.8 thus is likely to be
caused by competitive polymerization in the monomer phase; at high M/W due
to the presence of a mixed emulsion, and al low M/W resulting from a low
particle number. Yet, an additional factor may be that for Ihese large pattic1c~
n is possibly close to 05 (see chapter 7). Seeded experiments with monodisperse
latexes in this size range are intended to clarify this behaviour.
86 Chapter 6
6.4 Conclusions
Agitation needs to be sufficiently intense to prepare homogeneous
emulsifier-stabilized emulsions of butadiene, which is a prerequisite for
reproducible, gravimetric sampling (e.g. for conveaion). An unavoidable side
effect is the significant decrease in [El.1I> as a result of the small average droplet
diameter obtained under these conditions.
The large density difference D.p between butadiene and water promotes the
breaking of droplets in a turbulent flow field. Paradoxically, if turbulence is
insufficient ::'p stimulates (partial) phase separation, leading to a diffusion
controlled polymerization, whereas other monomers may still behave normally.
lncreasing M/W at constant [E] leads to a decrease in the final N as [El.ff
decreases_ An additional reduction of N was found at. M/W ::: 1, presumably
because mixed emul$ions are formed which reduce [E]d/ even further. The
presence of mixed emulsions in a polymerizing system is likely to affect the
kinetics, as polymerization in the monomer phase may no longer be negligible.
In fact, it was frequently observed that polymeri;>;ations at high M/W had two
kinetic regimes in interval It There are strong indications that the (change in)
type of emulsion is responsible for the anomalous kinetic behaviour at high
monomer /water ratios.
References 87
c.P. Evans, P.M. Hay, L Marker, R.W. Murray, O.J. Sweeting, 1. Appl. Polyrn. Sci. 5,39 (1961)
2 S. Omi, Y. Shiraishi, H. Sato, H. Kubota, 1 Chern. Eng. Japan 2, 64 (1969)
3 M. Nomura. M. Hamda, W. Eguchi, S. Nagata, I. Appl. Polym. Sci. 16, 835 (1972)
4 1'. Vermeulen, G.M. Williams, G.E. Langlois. Chern EI/g. Prog. 51, 85F (1955)
5 J.W. van Heuven, WJ. Beek, 1 Solvent &tr., proc. Int. Solvent Extf. Con! 1971, paper 51, 70 (1971)
6 W.A. Rodger. V.G. Trice, l.H. Rushton, Chern. Eng. Pmgr. 52,515 (1956) 7 R. Shinnar, 1.M. Church, Ind. Eng. Chern. 52, 253 (1960) 8 AH.l'. Skelland, G.G. Ramsay, Ind. Eng. Chern. Res, 26, 77 (1987) 9 S. Polslein, "Butadiene", in ''Monomm'', edited by E.R. Blout, w.l'.
Hohen$tein, RF. Mark, Interscience f>ublishers Inc., New York 1949 10 D.P. Durbin, M.S. El-Aasser, O.W. Poehleill, 1.W Vanderhoff, 1 Appl.
Polym. Sci. 24, 703 (1979) 11 H.W.H. Yang,! Macromol. Sci. - Phys. 820. 199 (1981) 12 W.V. Smith, RB. Ewart,l. Chem. Phys. 16,592 (1948) 13 S.M. Ahmed, M.S. El-Aasser, F.l. Micale, C.W. Poehlein, J.W. Vanderhoff,
''Rapid Measurernent of Adsorption Isotherms of Emulsifiers on Latex Particles", in "Polymer Colloids Ir, edited by R.M. Fitch. Plenum Ptes~, New York 1980, p. 283
14 K. Ishigure, T. O'Neill, E.P. Stahel, V. Stannett,l Mucrornol. Sci. - Chern. AS, 353 (1974)
15 M. Morton, P.P. Saiatiello, R. Landfleld,l. Polym Sci. 8, 111 (1951) 16 M. Morton, P.P. SalatielIo, 1 Polyrn. Sci 6, 225 17 K. Wendler, N. Karim, M. Fedtke, Plast.: Kautsch 30, 247 (l983) 18 M. Morton, P.P. Saiatiello, H. Landfield,l Polym. Sci. 8, 215 (1951) 19 LR. Howland, "GR·S Latex", in: "Synthetic Rubber", edited by G.S. Whitby,
c.c. Davis, R.F. Dunbrook, John Wiley & Sons Inc., New York 1954, p. 652" 655
Kinetics of the Emulsion Polymerization of Butadiene 89
Chapter 7 Kinetics of the Emulsion Polymerization of Butadiene
SUMMARY: Kinetic information on particle growth is obtained for the ab initio emulsion polymerization of butadiene. It is shown that the decrease in particle number by coagulation (induced by the high ionic strength) hardly affects polymerization rate, since the average number of radicals per particle (n) increases with particle size. From a rateanalysis of intervals U and III it follows that the system is 'zero-one' (n < 0.5), i.e. termination is not rate-determining. Zero-one kinetics, in combination with a low initiator efficiency, explains the small effect of initiator concentration on polymerization rate_ The radkal loss mechanism responsible for the zero·one kinetics could not be established unambiguously. but chain transfer/desorption processes involving polybutadiene, thiol, emulsifier, and the Diels-Alder dimer 4-vinyl-tcyclohexene, were all refuted on experimental groundS. Oesorption of monomeric species seems a reasonable explanation, given the rather low estimate of the propagation rate coefficient.
7.1 Introduction
Despite the enormous industrial importance of polybutadiene-conta.ining
polymers prepared by emulsion polymerization (c_g_ synthetic rubbers, high
impact materia.ls, coatings, adhesivcs). very little is known about the kinetic
behaviour of butadiene in emulsion polymerization systems.
As early as the synthetic rubber program unusual effects were observcc,l in
the (co)polymerization of butadiene(75)-styrene(25) or butadiene alone, but no
adequate explanations were provided_ These effects included the observations
90 Chapter 7
that trace amounts of thioh of low water solubility seemed to be e~sential to
bring about reaction in persulfate·initiatcd polymerh:ations,l) while the ratc of
(co)polymerization was unaffected when changing the concentration of persulfate
within wide limits.2)
In the previous chapters was discusscd the ab initio emulsion polymerization
of butadiene under conditions typically found in industrial practice, e_g_ high
ionic strength, large monomer to watcr ratios, and the use of a commercial
emulsifier. Although these experiments wefe not especially designed to study the
kinetics, they do contain a lot of useful information, which is now used to
an,llysc the typical behaviour of this mOnomer.
One of the principal problems encountered in interpreting the kinetics of
butadiene is the chronic lack of reliable rate coefficients in open literature. The
only values for the propagation rate coefficient kp stem from the early work of
Morton et al.,3) and at best represent an approximation of its true value since
they were determined under dubious experimental conditions (see chapter 6).
Within the limitations mentioned above an analysis is given of the kinetics
and mechanisms of the emulsion polymerization of butadiene, based on
experimental evidence discussed in the preceding chapters.
7.2 Interval II analysis
The overall rate of polymerization R"", (Lmol-l.s·1) is u::;ually expressed as
dc ill nM " ~ e", (n/N.,) N (7.1)
where n,/ is the number of moles of monOmer initially present per unit volume
of water, Ii the <lverage number of radicals per particle, NM Avogadro's number,
e", the monomer concentration within the particles, and N the particle number_
Since ~ is not known with sufficient accuracy, it is better to use the average
rate l)er particle (Rpol/N) as a semi-quantitative equivalent of n, because I), and
Kinetics of the Emulsion Polymerization of Butadiene 91
C" are COnStant in interval II to within an excellent approximation (see below).
Data can be conveniently presented as Jog-log plots of Rpot/N vs. partide
diameter d, in analogy with the well-known Ugelstad plots of log n v~. log , •. 4)
Figure 7.1 shoW5 Rpol/N vs. diameter for two polymerizations discussed earlier:
one with 16.2 g.L·1 dresinate 214 depicted in Figure 4.8, and with 8.1 g.L· l SDS
shown in Figure 5.2. Obviously, the observed decrease in particle number by
coagulation is compensated by an increase of n of the resulting particles, so that
1\01 remains virtually unchanged during the polymerization.
~ 60 ::;:: ~ 4.0 a;:"-f'j'
5: 2.0
1.0
0.6
OA
0.2
0.' +---r--.---....... --,.--,---,.....,...-r--~ 20 40 60 100
Diameter in nm
Figure 7-1. Average rate per patticle Rpc/N vs. diameter; dw (open symbols) and d. (full symbols)! for reactions with 16.2 gL dresinQle 214 (0), and Ii 1 g.L'/ SDS (~,.J-
Calculation of Rpo/N using ~I in interval II and the final particle number
after cessation of coagulation gives an upper limit of this quantity, since at lower
conversions RpoJN is evidently smaller due to the smaller particle size. Besides
type and COncelltration of emulsifier, other recipe parameters such as monomer I water ratio, ionic strength and initiator concentration were utilised to vary the
final particle si:1;e. Plotting the ave.age rate per particle on a log-log scale as a
92 Chapter 7
function of final particle diameter at an arbitrarily chosen final cunversion, viz.
9() %') (Figure 7.2 and 7.3), clearly shows that Rp<l/N depends solely on particle
si~e. Pulymerization kinetics is internally oon~i~tent, and otherwise unaffected
by variable~ such as emulsifier concentration and ionic strength, except in regard
to how these affect the final particle size. Furthermore, the general trend is not
affected by the method of particle sizing. Figure 7.3 also contains (lata points
taken from the work of Wendler et a1.S) for butadiene emulsion polymerizations
at 70"C, where particle size was varied by changing the amount of bis(iso
propoxythiuc<\rhonyl)disulfane, a chain transfer agent which decreases colloidal
st<lbility by introducing pol<lr end-groups that reduce emulsifier adsorptilln"(.)
Agreement with our results is excellent, given the temperature difference of 8"C
-"1 (; E ~
Z
20
10
., ., .--~ .. -o "
/ " ~"0.6 ~
I
b • 40 /
.~ Q
2.0 / •
1.0 / .. 60 /
.-04 ~
/ 02 •
40 60100 200
Figure ?2. Averoge rate per particle Rp<>/N vs. the diameter dO' ar 90 % conversion (dooJ, for polymerizations with dresinate 214 (open symbols) and sodium dodec-ylsuifate (full symbols) . The final particle size was varied by changing rhe emu.lsifier concentration Ct (0), cation concentration (6.). o(
monomer/water ratio MjW ('\1) in Ihe standard recipes.
400
.) 90 % Conversion W<lS cho~en for the sake of experimental convenience, as the majority of polymerizations was stopped between 85 and 95 %.
Kinetics of the Emulsion Polymerization of Butadiene 93
20
"6 .§. 10 ""'-Z a:."-e.o ~ ~ 4.0
2.0
1_0
0_6
0.4
40 60 100 200 400
Figure 7.3. Average rale per particle Rpo/N vs_ the diameter' dv at 90 % conversiof! (d~. Data of Wendler el aL 5) for polymerizations at 7(/'C (0), For other symbols see Figure 7.2.
Using the uncertain value for ~ "" 100 Lmor1.s,l reported by Morton3) at
62"C, and C" '" 5.6 mol,L-! (see Appendb[), a value of 4.7 lO-u mo\.s-l for
~/N would result if n '" 05, Given this rough estimate of a Smith-Ewart case
2 situation, and again emphasizing that the calculated values for RI"'II'N are
upper limits for the experiments involved, it is evident from Figures 7.1 - 7.3 that
values for Ii < 0.5 are realistic.
Importantly, no constancy of Rp<>l/N associated with n '" 0.5 is observed in
the particle size range covered. The apparent leveling orf at c:4o ;;. 200 run may
be artificial, because the polymerizations involved are extremely slow, sometimes
taking 2 to 3 days to complete conversion. On such a time scale bulk polymeri·
zation in the mOnomer droplet~ can no longer be neglected as is usually done_
This behaviour clearly shows up in the particle size distribution of the final
latexes, which are extremely positively skewed (see Figure 6_7). Therefore these
experiments are ignored in the following diSCllssion.
94 Chapter 7
7.3 lnterval III analysis
The complications associated with particle nucleation and coagulation can
be avoided by using the kinetic information in interval III, where the particle
numher is truly con~tant (see Figures 4.8 and 5.2). Since monomer concentration
in the particles is decreasing continuously. it is convenient to remove CM from
the right· hand side of equation (7.1) and write the expression in terms of the
fractional conversion in interval HI, x. with )( = (C,,,o " CM)/C".t and C,.." the
initial monomer conGentration in the particles:1)
dln(l - x)
dt (7.2)
where il ... is the number of moles of monomer per unit volume of water present
at the heginning of interval Ill. Since nM '" (1 - x')n"..", with x' the conversion
where interval III begins, we can Galculate x' from the experimental values of
the rates in interval n (R",,]) and interval III (-dln(J . x)/dt). The Gonversion x'
is found to he 0.60 (standard deviation == 2.3 %, using 25 data points), and
independent of particle size (Figure 7.4). This latter observation further
substantiates the common assumption of Gonstancy of CM in interval n.
oF- 70
.5 c 0 '~ aJ 60 > c 0
o o_o_oo~~ __ ~ ___ o __
o o o 0 o 0
50 0 30 60 90 120 i 50
Diameter in nm
Figure 7.4. The fractional conversion where interval III begins (x') VS.
panicle diameter d", at .r'. An average value x' ;;= 0.60 wa.~ obtained.
Kinetics Qf the Emulsion Polymerization of Butadiene 95
The value of x' is somewhat higher than those reported in literature based
on vapour pressure measurements, viz. 0.528) and 0.56.3) These latter values are
considered more inaccurate because the non-ideal gas behaviour of butadiene
vapour complicates an accurate detennination of the conversion at the point
where pressure begins to drop.
For the simple zero-one system (i.e. only particles with zero or one free
radical per particle need to be considered) where bimolecular termination is not
rate determining, it is easily shown that (see section 2.2):9)
ii (2.7)
where p is the rate coefficient for radical entry and k the rate coefficient for exit
(desorption). During interval III, p may be considered unchanging to within a
good approximation, since the swollen particle diameter is almost constant. On
the other hand, k may depend on factors other than the particle volume, e.g. the
monomer concentration e",. However, a reasonable starting point is to assume
k to be constant, and then to check if this is consistent with the data. Several
limiting cases can be distinguished:
Smith-Ewart case Z, with k « p:
Smith-Ewart case 1, with k » p:
ii '" 0.5
ii "" pfk
Plotting -In(l - x) vs. reaction time (Figures 7.5 and 7.6) gives straight lines
up to a weight fraction of polymer wp (equivalent to the ab initia conversion c)
of 0.85 or higher, for all polymerizations with <40 < 175 nm.
Since the bimolecular termination rate coefficient varies significantly with
wI>' tennination cannot be rate-determining because -dln(l - x)fdt is constant for
0.60 < wp < 0.85 under all conditions investigated. The approximation of
instantaneous termination is thus valid, and n < 0.5.
96 Chapter 7
.,,1.5 <) "
:=.. / ",/ 0 ,,/ .5 I
0 / ,,/ I ,,/
1.0 " / II / / / " / ,,/ / c
<) / ,/ I /
0
./ 0 / 0.5 j / 0 ,,/
/ / / v 0
/ I' / / ./ " .. " 0.0
0 2 4 6 S 10 Time in hours
,,1.5 0
0.9 I ID ~ " E
I ,,/ "" l/ "0
/ 0.
'0 1.0
o v J,/ c 1/ c / / .Q
1/ c
/,/ g
/ c ~
/ 0.8 E / " /' ./ .g>
/1 /' ., ,./ :s:
0.5
//,,// ,,/ .. /
gV,,/,,/" 0.7 /,,/
:%q,/ a:/
0.0 0.6 0 2 4 6 8
Time in hours
Figure 7.5 (top). "In(l" x) vs. Reaction time in inrerval JJI, for po/rmen:atiol1S wi~? diff.erenl !JJS conce:ztrat~ons c:.." '" 64·~f"L-l (0), C. - 32.4 g.L (v), CE - 16.2 g.L (D), C" - 8.1 gL (,e,.). For convenien(:t~ the zero·point is shifted one hour along the x"axis.
figuft' 7.6 (bottom). -In(1 - x) vs. Reaction time in interval III, for polymerizations with different dresinate 214 concentrations CF. = 64.8 g.L·l (0); C. '" 32.4 gLl (<;J); C. = 16.2 gLl (D); C ... 8.1 gL· l
(/',.). Orr the right axis is indicated the weight fraction oJ polymer wp
Kinetics of the Emulsion Polymerization of Butadiene 97
Furthermore, Figures 7.5 and 7.6 show that the product of 1;.. ii is constant
within each experiment; it would be highly fortituous if k,. and Ii would
counterbalance under all experimental conditions, so it is apparent that"" and
ii are both constant in inteIVal III. This further implies that k., is not diffusion
controlled, at least for wp <;; 0.85; this is to be expected given the low glass
transition temperature of (emulsion) polybutadiene (Tg = - 866C).IO)
The termination rate coefficient may start to become sufficiently small for
very large particles as to become rate-determining, consistent with the non
linearity of -In(l - x) observed for polymerizations with d90 > 175 nm (e.g. the
polymerization with 8.1 g.L·' dresinate 214 in Figure 7.6, with d.. = 180 nm at
90 % conversion). At a relatively low w~ the interval III rate -dln(l - x)/dt, and
thus 'ii, begin to increase. Termination is no longer instantaneous, yet the initial
steady state suggests a zero-one behaviour at lower wp' If we assume the initial
jj '" 0.5, an estimation for k,. = 180 - 200 L.mol-l.s-l is obtained. Although
intuitively not unreasonable this assumption is not necessarily correct, since the
initial n can still be < 0.5. The particle si:>;e distributions (PSD) of these latexes
are quite narrow, but not monodisperse, which further compliciltes a proper
interpretation. Therefore this estimate of I), should be regarded as a lower limit
of its true value.
7.4 Effect or the initiator concentration
Having established that the studied system shows zero-one behaviour Over
a wide range of experimental conditions, we can further explain the effect of
initiator concentr~tion [l] on polymerization kinetics (Table 7.1). The initiator
concentration was varied lOO-fold between 26.0 and 0.26 mmoLC1, while every
change of [I] was accompanied by an appropriate adjustment of the amount of
K+ INa+ carbonate to keep the overall cation concentration constant. Coagula
tion phenomena are thus a constant factor in these experiments.
Under all conditions tried, and in agreement with previously reported
results,2) both Rpol and -dln(l - x)/dt are only weakly dependent On [[J.
98 Chapter 1
Table 11. Initiator exponents with regard to parriele number (N "" [It) and polymerization rates in intel1Juls If (~I ~ [I]Y) and III (-dln( 1 - x)/ dt '" [11Z~ at different emulsifier concentrations C£ and cation concentration.~ [cation}.
Recipe CB [cation] L- l motel g.
2 32A 0.3 2 16.2 0.3 1 32.4 OA . ) ~ '" number of e)(periments.
20 ~ ., 0 E
10 , 2: rr:"" 6.0 " 'b
4,0
2_0
1.0
0_6
0.4
0.2
/ o
I o I
1
,. ,. i ~
/ o I o
40 60 100 200
~.)
9 7 7
d9()b) )( y z
run
0.46 0.04 0.04 46 0.16 0.09 0.09 80 0.02 0.08 94
b) ~ at [I] '" 13.0 mmoI.L-i .
Figure 7.1 Average rate per partide Rp<>IN vs. the diameter d .. at 90 % conver:rion (d~, for polymerizations with a lOO-/old variation in [Sp/} (see Table 7.1) at nvo concentrations SDS (recipe 2) C~ '" 324 S·L-! (0); CE '" ]6.2 g.L'i (e). The solid line represents the empirical relationship from Figure 7.2.
400
The average rate per particle as a function of particle size for the first two
series of experiment~ (recipe 2) from Table 7_1 are shown in Figure 7.7, where
the solid line represents the best fit of data points from Figure 7.2. Again the
Kinetics of tbe Emulsion Polymerization of Butadiene 99
kineti(:s is internally consistent, yet the l00-fold variation of II] has an almost
neglihle effe(:t on kinetics. This implies that p is independent of II] in the
present system, i,e. the initiator efficiency must be extremely low. Furthermore>
for latexes with a narrow PSD. p can be taken inversely proportional to N.
Substituting this into equations (7-1) and (7.2) and assuming a low value for n, viz. ii '" p/k, render both ~l and -dln(l - x)/dt independent of (I] and N. This
is in good agreement with the data in Table 7.1, especially for the first series of
experiments, since the assumption of ii « O.S is most valid for small particles.
7.S Radical loss mechanisms
It is obvious that a first-order radi(:al loss process is kinetically dominant in
this system, whose nature is as yet uncertain. Several possibilities may arise:
(1) First-order termination by monomer occlusion is highly unlikely, since
monomer and polymer are completely miscible (or highly 5wellable if the
polymer is l;fQss-linked), while the glass transition temperature (Tg -- _ 86"C) is
well below the reaction temperature of 62"C.
(2) Radical trapping by transfer to polybutadiene, as proposed by Hagiopol
et al. 11) for the emulsion copolymerization of styrene and acrylonitrile in the
presen<:e of PB, is not important, sinl;e ·dln(l - x)/dt (and thus n) is constant
over a period of several hours, whereas wp changes signifil;antly during this time.
(3) Chain transfer to thiol and subsequent desorption of thiol radicals I;3n
also be ruled out, since Nomura et aI. 12) showed that n-dodecanethiol radic:als
do not desorb because of their extremely low water solubility. Furthermore,
omitting the Ihiol from the tel;ipes used in this investigation never increased
Rpol/N (see chapter 8). as would be expected in case of desorption of thiol
radicals. Instead, a marked de(:rease of Rpol/N was sometimes observed.
100 Chapter 7
(4) It is well-known that butadiene ea.sily undergoes Die1s-A1der cyc\izations
at elevated temperatures, giving products such as 4-vinyl-1-cyclohcxene (VCH).
Although it is described as a mild retarder in the copolymerization of
butadiene(75)-styrene(25)/3) addition of 1 wt.-% VCH to a butadiene emulsion
polymerization using SDS (recipe 2) had no effect whatsoever, either on the
conversion-time history or on the final particle number. Furthermore, addition
up to 4 wt.-% YCl-l to a styrene emulsion polymerization (using recipe 2 without
thiol at 50"C) did not a.ffect the molecular weight significantly (Table 7.2).
Although inherently present during polymerization, this Dicls-Alder dimer does
not seem to facilitate the extensive radical desorption observed experimentally.
Tuble 7.2. Numher-average (M.J and weight-average (M,.) molecular weight detennined by GPC of polystyrene samples, prepared according to recipe :;ti
tit S(fC, without t-dodecanethiol (t-DT) and different amounts of VCH
yew) 0% 0.25 % J% 2% 4% 1 % t-DT')
1O-5.Mn 4,1 4.3 3.9 3.8 3.9 0.49
lO-6.M" 2.1 2.5 2.3 1.8 1.7 0.21
a) weight percentage of VCl·-j and t-DT on monomer basis, b) [I] = 13.0 mmol.L-1 and C" "" 16.2 g.t;1 SDS.
(5) Many common emulsifier, (including sodium dodecyl~ulfate) can act as
chain transfer agents. 14) However, this phenomenon cannot be of major
importance in the present system, since polymerizations yielding the same final
particle size have the same value for Rpol/N, regardless of the type of emulsifier.
(6) Lastly, transfer to monomer (or dead oligomeric material), followed by
exit ()f the monomeric radical seems reasonable given the rather low estimate
of kp• In the transfer/diffusion model developed by Nomura anu Harada15•16) a
propagating polymer chain transfers its free-radical activity to a monomer
molecule (or to a chain transfer agent), which then diffuses to the particle
surface where it desorbs. The process is completed when the radical diffuses
aw,ly from the surface into the b\llk aqueolls phase. When all of these three
Kinetics of the Emulsion Polymerization or Butadiene 101
sequential events are significant, the exit rate coefficient k is given by:
k 122; D.q k"
(q + ZD.iDp)d,.,2 ~ (73)
where D.'1 and Dp are the diffusion coefficients of the exiting species in the
aqueous phase and in the particle, respectively, :;l is roughly equal to the
maximum degree of polymerization of the exiting radical,b) dow the swollen
particle diameter, and k" and kp the rate coefficients for transfer to mOnomer
and for propagation, respectively. The parameter q is the partition coefficient of
the exiting species between the particles and the aqueous phase, and is roughly
equal to the ratio of saturated monomer Solubilities in the particles and aqueous
phase (C",/C.q). A similar expression was derived by Hansen and Ugelstad. l7)
If in the transfer/diffusion mechanism for radical el(it transfer to monomer
is the rate-determining step, equation (7.3) can be replaced by:IS)
(7.4)
Obviously, this is not the case in the present system, since ii, and thus k,
were found constant fa. into interval Ill, whereas c.. decreases continuously.
However, the constancy of k is consistent with the general equation (7.3), since
d.w and q are not expected to vary significantly in interval m. For a sparingly water.soluble monomer q » 2D.,/Dp' so that equation (7.3)
can be approximated by:18)
(75)
b) The parameter z '" E[~CM/(kn + ts,C ... )P. where k is a desorption (ate coefficient, and j the degree of polymerization of the exiting species. with 1 ::: j ::: j ..... When ~C .. » kn, this gives z ::. jrn •• '
102 Chapter 7
Table 7.3. Kinetic data for butadiene and styrene at 5(J'G.
Parameter Butadiene Ref. Styrene Ref.
kir/L.morl.s"1 7.W' 19) ~/Lmorl.s"1 60 3) 258 20)
&q/mmo1.L"1 6 2J) 3.7 22) 37 23)
C ... /mol.L· t 5.7') 5.8 9)
.) Estimated from eM determined at 62"C in the present study.
On the basis of equation (7.5) an estimate can be made of the order of
magnitude of k for butadiene, relative to the one for styrene at the same dow.
Values for e .. , Ca<j and 10), for both monomers are given in Table 7.3.
According to the Wilke-Chang correlation21) the diffusion coefficient D.q b
related to the molar volume of the solute at its boiling temperature under
atmospheric pressure, V"" according to D.q '" V,; 0.6. It follows that D.q for
butadiene is about 1.3 times larger than the one for styrene.
Ba~ed on chemical structure, the maximum degree of polymerization of the
cJ>iting species z is expected to be roughly two times larger for butadiene (ClI6)
as compared with styrene (C8H~).
Unfurtunately, no value for k" has been reported fur butadiene. However,
values for k" and kp are known for 2-chloro-l,3-butadienc at 50·C, determined
from low conver~ion solution polymerizations, viz. kit = O.18 L.mor1.s·1 and 10),
co 792 L.morl.s"I?~) Initiator efficiency was found to be 0.4 and the average
degrees of polymerization between 2000 and 3000; both quite normal values for
solution ~xperiments, indicating a good stability of the chlorine a.tom towards
chain transfer. For styrene ki,/"r between 0.35 and 0.78 10" have been reported,
while for o-ehlorostyrene a similar value is found, viz. 0.27 10"'\ all at 50"c.26)
Apparently, the prc~ence of a chlorine atom hardly affects kir; thus for butadiene
ktr ", 0.1 LnlOrl.~.1 at 50T seem..~ quite reasonable.
There appears to be some confusion about CM for butadiene. Most values
reported (and referred to thereafter) were determined at 2S"C and 1 atmosphere
(eM ",,14 - 15 mmoI.L-1),2.27.:lS) while polymerh:ations are normally performed at
saturation pressurc. This markedly increases C", as is obvious from Figure 7.8.
Kinetics of the Emulsion Polymerization of Butadiene 103
"'" 60 • ;.,
. / 'ci :2 50 d b 40 ./ 0
\ 30
0
20 \ 10 O~
0,-o~o
0 ·20 0 20 40 60 80 100 120
Temperature in "C
Figure 7.8. Solubility of butadiene in water Caq as a function of temperature at 1 atmosphere (0),21) atld at butadiene saturation pressure (e)- 23)
Based on the above estimates for the parameters in equation (7.5), and
using C.q .. 37 mmol.l-1 at SO"C, the exit rate CQefficient for butadiene is
expected to be at least one, and possibly two orders of magnitude larger than
the one for styrene. Altogether, radical desorption by monomeric species in the
emulsion polymerization of butadiene appears reasonable, but experimental
evidence is as yet insufficient to substantiate this possibility. The similarity
between the emulsion polymerization of more water-soluble monomers (e.g.
vinyl acetate, methyl acrylate), kinetically dominated by desorption of
monomeric radicals, and the butadiene emulsion polymerization, suggests that
similar events determine the kinetic course in the present system.
Further research is necessary to elucidate the exact kinetic scheme, but the
new results presented here certainly provide a much better understanding of the
behaviour of this widely used monomer in emulsion polymerization systems.
104 Chapter 7
7.6 Conclusions
The average rate per particle in interval II was found to depend strongly on
particle size, indicating that ii is either smaller or larger than 0.5. No constancy
of Rpol/N associated with Ii = 0.5 was found in the particle size range covered.
An analysis of interval III rates showed unambiguously that the present
system follows zero-one kinetics with Ii < 0.5, so that second-order termination
cannot he rate-determining (except perhaps for very large particles).
By combining interval II and interval III rates of a single e1l':periment, the
conver~ion where interval III begin5 could be calculated. After correction for
the non-ideal mixing of monomer and polymer, the saturation solubility of
butadiene in polybutadiene was obtained, and found constant in interval II and
independent of particle size.
Over the range of particle number N and initiator concentration [I] studied,
RpoJN (and thus p) was found only weakly dependent on [I], indicating a low
initiator efficien<,:y. This, in combination with the established zero"one kinetics
of the system, makes Rpol virtually independent of [I] and N, in conconlance
with experimental findings.
Loss of free-radical activity within the particles is a first-order process, which
is likely to he transfer to monomeT followed by exit. The desorption rate
coefficient of butadiene is expected to be significantly larger than the one of
styrene, where ii is often found close to 0.5 within fairly wide experimental
limit~. This al5o, at least qualitatively explains the observation that upon
increasing the initial butadkne/styrene ratio from zero to one in the emul5ion
copolymerization of these monomers, the p<J.rticle number increases, while tne
polymerization rate decreases (see section 1.1 ).29,30) Beside a decrease in
<J.pparent I), as the fraction of butadiene increases, i'i is also expected to drop off
$ignificantly, ~~ a result of an increased exit rate.
Kinetks of the Emulsion Polymerization of 8utadiene 105
Appendix Calculation of the saturation monomer concentration
The beginning of interval III at conversion x' corresponds to a situation were
the unreacted mOnOmer is almost completely imbibed in tile latex particles. For
butadiene the small amount dissolved in the aqueous phase can be ignored.
Tak:ing conversion x' at 0.60 (see section 7.3) and using Pp '" 0.88 g.mL·1 and
fJ .. = 0.565 g.mL-1 at 6TC, a value for e" '" 5.32 mol.L·1 is obtained, assuming
ideal mixing of monomer and polymer. However, systems like these rarely
behave ideally, which is revealed by measurements of the density of the
polymerization system as a function of conversion. For this purpose an Anton
Paar density meter with a remote cell (DMA 401, max. pressure 10 atm.) was
employed. Samples were taken at regular time intervals, in between which the
cell was flushed with distilled water. Conversions Wel"e determined gravimetri
cally as described previously (see section 3.2).
~ 097 E
!;; ~ .~ 0.93
'" C
0.89 Q"
.. '"
O. 81+----.---,--,---------r~_,_-___.-_._-.___"_,,________, o 20 40 60 80 100
Conversion in %
Figure 7. 9. Density of the reaction mature vs. gravimetric conversion. Symbols represent experimental points from three polymerizations (recipe J). and the solid line the theoretical relationship, assuming ideal mixing of mOnOmer and polymer.
106 Cbapter 7
The experimentally determined density-conversion relationship (Figure 7.9)
differs markedly from that calculated assuming ideal mixing of monomer and
polymer and additivity of specific volumes of the separate phases. Mutual
solubility of butadiene and water is low enough to be ignored.
At conversions < 50 % the di~erepancy between ideality and experiment
may be partly caused by the instability of the monomer emulsion. Demixing
started seconds after the cell was filled with a fresh latex sample. In the absence
of a steady signal the minimum value for the density was taken, which may
introduce a serious systematic error. It should be noted though that this
minimum value was quite reproducible. In contrast, at conversions > 50 %
measurement~ were free of this artifact and can be used without reservations.
At conversion x' = 0.60 only tbe aqueous phase and the monomerrswollen
latex particles have to be considered. Additivity of tbe specific volumes of both
phases is justified, since mutual solubility is negligible. From the experimentally
determined den~ity at 60 % conversion, the density of the swollen particles can
be calculated. A volume contraction factor then can be defined as the ratio of
'ideal' to measured density of the swollen particles, which was found to be 0.95.
The saturation monomer concentration C" corrected for non-ideal mixing of
monomer and polymer thus becomes 5.6 mol.L-1, and was found independent of
(unswollen) particle diameters between 30 and 150 nm (Figure 7.4).
References t07
1 LM. Kolthoff, W.E. Hanis, J. Polym. Sci. 2,41 (1947) 2 F.A. Bovey, I.M. Kolthoff, A.I. Medalia, EJ. Meehan, In ''Emulsion
Polymerization", Intersdence Publishen:;, New York 1955 3 M. Morton, P.P. Salatiello, H. Landfield,l Polym. Sci 8, 215 (1952) 4 ], Ugelstad, P.e. Mark, J.~. Aasen, J. Polym. Sci.; Part A·] 5, 2281 (1967) 5 K. Wendler, N. Karim, M. Fedtke, Plaste I<mttsch- 30, 247 (1983) 6 K. Wendler, H. Elsner, W.D. Hergeth, M. Fedtke, Plaste KaUlsch. 32, 128
(1985) 7 D.R. James, D.C. Sundberg'/' Polym. Sci., Polym. Chem. Ed. 18, 903 (1980) 8 ET Meehan, J. Am. Chern. Soc. 71, 628 (1949) 9 R.G. Gilbert, D.H. Napper,], MacromoL Sci., Rev. Macrornol- Chern. Phys.
C23, 127 (1983) 10 L,A. Wood, ''Physical Chemistry of Synthetic Rubbers", in "Synthetic Rubber",
edited by G.S. Whitby, c.c. Davis and R.F. Dunbrook, John Wiley & Sons, New York 1954, p.358
11 C. Hagiopol, T. Deleanu, T. Memetea, J. AppL Polym. Sci. 37, 947 (1989) 12 M. Nomura, Y. Minamino, K. Fujita, M. Harada, J- Polym. Sci., Polym_
Chern. Ed. 20, 1261 (1982) 13 WX Taft, GJ. Tiger, In "Synthetic Rubber", edited by G.S. Whitby, c.c.
Davis, R.F_ Dunbrook, John Wiley & Sons, New York 1954, p. 683 14 I. Piirma, V_R_ Kamath, M. Morton, J. Polym. Sci., Polym. Chern. Ed. 13,
2087 (1975) 15 M. Nomura, M. Harada, J. Appl. Polym. Sci 26, 17 (1981) 16 M. Nomura, "Desorption and Reabsorption of Free Radicals in Emu-lsion
Polymerization", in "EmulsionPolymerkation". edited by I. Piirma, Academic Press, New York 1982, p. 191
17 F. K. Hansen, J. Ugelstad, MakromoL Chern. 180,2423 (1979) 18 M. Adams, D.H. Napper, R.O. Gilbert, D.E Sangster, 1 Chern. Soc.,
Faraday Trans. 1 82, 1979 (1986) 19 AW. Hui, A.E. Hamielec,1 AppL Polym. Sci. 16,749 (1972) 20 S.W.Lansdowne, R.G. Gilbert, D.H. Napper, D.F. Sangster,/. Chem. Soc.,
Faraday TrCl1!S. /76, 1344 (1980) . 21 W.M. Salt man, ''Butadiene Polymers", in ''Encyclopedia of Polymer Science
atui Technology", Inten:;cience, New York 1965, 1st ed., vol. 2 22 FA Bovey, I.M. Kolthoff, J. Po/ym. Sci. S, 487 (1950); mentioned in 2)
23 cn. Reed, JJ. Mcl(etta, J. Chern. Eng. Data 4, 294 (1959) 24 c.R. Wilke, P. Chang, Am. Inst. Chern, Eng. J. 1,264 (1955) 25 F. Hrabak, M. Bezdek, Collect. Czech. Chern. Commun. 33,278 (1968) 26 LJ. Young, 'TrCl1!Sfer Constants to Monomer, Polymer, Catalyst and Solvent
in Free Radical Polymerization", in "Polymer Handbook", J. Brandrup and RH. Immergut, Wiley.lnterscience, New York 1975, 2nd edition, p. 11-61
27 C. McAuliffe, J. Phys. Chern. 70, 1267 (1966) 28 S. Banerjee, Environ. Sci. TechnoL 19, 369 (1985) 29 RS. Bhakuni, Ph. D. Thesis, University of Akron (1964); Chem_ Abstr. 62,
13351b (1965) 30 V.l. Eliseeva, S,S. lvanchev, S.1. Kuchanov, A.V. Lebedev, "Emulsion
Polymerization ami. its Applications in Industry", Khimiya, Moscow 1976; English Trans!.: S.l. Teague, Consultants Bureau, New York 1981
IUl'eds or Thiols 109
Chapter 8 Effects of Thiols
SUMMARY: The role of thiols of low water solubility, commonly uSed in the emulsion polymerization of butadiene, has been investigated. The following effects have become apparent: (1) these thiols act as efficient chain transfer agents in limiting the formation of heavily cross-linked polymer networks; (Z) the monomer concentration within the particles is not affected by the thiol; (3) thiol radicals do not desorb because of their extremely low water solubility. The socalled 'promoting effect' of thiols in emulsion polymed7.ations of diene-hydrocarbons in fact appears to be related to impurities present in the emulsifier; this effect is completely absent in emulsifier-free polymerizations.
8.1 lntroduction
In earlier chapters the ab initio emulsion polymerizations of butadiene with
dresinate 214 and sodium dodecylsulfate as emulsifiers, both in the presence of
tertiary dodecancthiOl as chain transfer agent, were discussed. Chain transfer
agents are routinally used in diene-polymerizations to limit the extent of
branching and cross-linking of the polymer, and thus the amount of heavily
cross-linked insoluble gel. A limited gel content greatly facilitates processability
of the crude polymer, such as in the compounding of synthetic rubbers.
However, chain transfer agents also ellert other, sometimes less desirable
effects in emulsion polymerization systems. It is well known they can promote
radical desorption, t.2) thus lowering the rale of polymerization. Furthermore,
lIO Chapter 8
slightly water-soluble thiob were found to be essential to b(ing about reaction
at acceptable rates in many persulfate-initiateo polymerizations of butadiene(75)·
~tyrene(25f) or butadiene,4) a phenomenon referred to as the 'promoting effect'.
This chapter describes the various effects of the chain transfer agent tertiary
dodecanethiol (t-DT) in the u/) initio emulsion polymerization of butadiene, with
regard to the process of gel formation as well as to the reaction kinetics.
8.,2 'J'hiol as chain transfer agent
The gel point (i.e. the conversion at which the first insoluble polymer
network makes its appearancc) was determined with the vistcx and extraction
proceoures described in section 3.6, for polymerizations with dresinate 214
(recipe 1) ao; emulsifier (Figure 8.1). Agreement between these techniques is
good, as in both cases the same solvent/diluent was used. Results with varying
amounts of t-DT arC summarized in Table 8.1.
Figure 8.1. Detennination of the gel poim and gel colltent with the viste;r; (0) and toluene extraction procedure (er
'" 200 OJ --i 5 180 ~
.,. 160
140
120
100 60
(~\ / 80 <f.
~
1: 60 Ql
1": o \j. 8
I 1j
40 &
:20 / 0
0 • 0
70 80 90 100
Conversion in %
Effects of Thiols
Table 8.1. The gel points for polymerizations with dresinate 214 (recipe 1) and different amOunts of t-dodecanethiol (t-DT).·)
gelpoint in % % conversion after 11Y'~1 g.g"1 visteJ! extraction 2hr 4hr 6hr 8hr g.mL-'.s-1
0-0 25,26 28 16 42 69 85 1.6 3.5 68,73 72 l8 42 68 87 1.6 7.0 83,84 83,84 15 43 71 89 1.7
0) Particle number is l.0 (± 0.05) lOIS mL'! in all experiments_ b) C'.DT expressed in grams t-DT per gram of butadjene.
111
As expected, the gel point goe5 to higher conversions with increasing
amounts of thiol. However, there appear~ to be no significant effect on the
polymerization kinetics; both ~I and N are identical within experimental error.
Cross-linking and network formation may possibly affect (he monOmer
solubility within the latex particles (and thus the kinetics). Even when monomer
and polymer are completely miscible, the swelling of a particle is limited,
because the free energy of mixing, which favours swelling, is counterbalanced by
the resulting increase in interfacial energy. The monomer concentration CM at
saturation swelling is described by the Morton-Kalzerman-A1tier equation:»
- Ii. - 1f.2 In(1 - 1/1i.) 4 V" "I I )( + If.S 3 d RT
(8.1)
where If. is the volume swelling ratio, V .. the molar volume of the monomer,
d the unswollen particle diameter, x the Flory-Huggins interaction parameter,
"I the interfacial tension, R the gas constant and T the absolute temperature.
Ii. is related to the monomer volume fraction ¢M by: If. = 1/(1 - 4>,,).
for densely cross-linked latex particles the contribution of the elastic energy
in limiting the swelling capacity is given by the F1ory-Rehner theory,6) and is
expressed in equation (8.1) by an extra term on the left-hand side:
(8.2)
112 Chapter 8
where Pp is the polymer density and M. the average molecular weight of the
segments between rwo cross-links.
Given thc large variation in gel point in the experiments described above,
and consequently the large variation in cross-link density, the effect on eM appears negligible. Apparently, even in the absence of t-DT the cross-link density
is still too low (or alternatively M. too large) to further restrict monomer
solubility within the particles under the prevailing experimental conditions.
In section 7.3 an average value of x' '" 0.60 (standard deviation 2.3 %) was
determined fot the conversion x' where interval III begins, and this value was
found to be independent of particle size between 30 and 150 nm. The values of
x' at different levels of I-DT for polymerizations with SDS (Table 8.2) and
dresinate 214 were also found identical within experimental error, and in perfect
agreement with those previous results. Consequently, the monomer concentration
eM must be constant and independent of the amount of t-DT.
Table 8.2. The conversion x' where interval 111 begins, for polymerizations with 32.4 gLI SDS and varying amounts of I-DT.
Hr.c,_DT!g·g" conversion x'
0.0 0.616
8.3 Promoting etred of thiol$
0.0 0.609
0.7 0.598
1.4 0.610
2.8 0.625
Another important effect of slightly water-soluble thiols is the socalled
'promoting effect'. At the time of the synthetic rubber program during World
War II it was already known that the persulfate-initiated emulsion polymeriza
tion of butadiene or butadiene(75)-styrene(25) was extremely slow in the absence
of an appropriate thiol. In the presence of such thiols, Rpol is virtually
independetlt of the persulfate concentration within wide limits/) and above a
Effects of Thiols 113
certain minimum level also independent of the amount of thiol.3) Similar results
were found for the emulsion polymerization of isoprene.G) This rate-enhancing
effect was o:;ommonly explained as being caused by a redox reao:;tion between
persulfate and thiol. Hydrogen abstraction by sulfate radical-anions would give
uncharged, lyophilic thiol radio:;als, which supposedly are more efficient in
entering a latex particle. Indeed, thiols containing 8 to 12 C-atoms solubilized
in emulsifier solutions are being oxidized by persulfate to disulfides?) The latter
o:;ompound is easily envisaged as being formed by combination of tbiol radicals.
The promoting effect is reo:;onsidered here qualitatively. using as emulsifiers
dresinate 214, SDS, potassium stearate and potassium oleate, and a~ initiators
sodium (SPS) or potas~ium persulfate (PPS), 4,4'-awbis(4-cyanopentanoic acid)
(ACPA) and 2,2'·azoisobutyronitrile (AIBN). In a preliminary communication4)
(Table 8.3) it was shown that the promoting effect was specific of persulfate
initiated polymerizations, with dresinate 214 as emulsifier. However, at lower
concentrations of ACPA, the polymerizations in the absence of t-DT also
become markedly slower (Figure 8.2, extension of Figure 4.1). The type and
amount of initiator clearly plays an important role.
r- 3.0 ."1
:...J E 2.0 0)
::::::: s. 0:; '0 ... 1.0
0.6
0.1 0.3 0.6 1.0 3.0 6.0 10 30
1o·,3.R/(g·'.s·') Figure 8.2. Variation of polymerization rate R~ witll radical ptoduction
rate per gram emulsion R for three initiators in the presence (open symbols) and absence (full symbols) of t-dodec(Jnethioi (recipe 1): PPS (0), ACPA
(1::.) and AIBN ('V). FOt details on R see section 4.2.
114 Chapter 8
The type of emulsifier also appears to be crucial in the promoting effect of
toOT, as Table 8.4 reveals large differences between the emulsifiers employed.
It may be noted that lshigure et al. IO) found a similar increase in R.,.,I of about
30 % with SDS in the radiation-induced emulsion polymerization of butadiene.
A~ pointed out by other investigators3) reproducibility of the experiments
without thiol is greatly affected by traces of impurities. Bhakunill) showed that
with peroxide-containing emulsifiers some polymerization occurs in the absence
of thiol, whereas with peroxide-free emulsifiers no polymerization was found_
'rtw results in Table 8.4 confirm the suggestion that impurities in the emulsifier
need to be considered in relation to the poorly understood promoting effecL
Altogether, the experimental data presented here qualitatively confirm the
observations from early literature, but also gross quantitative differences become
apparent, related to the nature and concentration of emulsifier and initiator.
This certainly is not helpful in clarifying the mechanistic role of thiols of low
water solubility in the emulsion polymeri:;::ation of butadiene. In an attempt to
eliminate artifacts introduced by the emulsifier, ab initio emulsifier-free emulsion
polymerizations with potassium persulfate as initiator were planned.
8.4 Emulsifier-rree polymerizations
Bmulsifier-free polymerizations in the presence and absence of t-DT were
performed; tne monomer/water ratio was chosen low (i,e, 1/9 by weight), to
reduce the polymeri:;::ation time to an acceptable level (after 25 hours a
conversion of 90 % was reached).
The kinetics in interval II of both emulsifier-free polymerizations, ilnd of
similar polymerizations with emulsifier and t-DT, were all found identical within
experimental error, taking into account the variation of Rpol/N with particle
size. 'Inc~c data are consistent with the observation thaL RpoJN is solely
determined by particle size, and not by the parameters utilised to manipulate the
final particle diameter (see section 7.2). This behaviour strongly suggests that the
promoting effect is indeed related to impurities present in the cmubifier.
Table 8.3. Effect of the type mut concemration of initiaJor on reaction kinetics, both in the presence and absence of t-DT. Fim part wi!h dresmrue 214 as emulsifier, and second pan with SDS (CE = 32.4 g.LJ
). t-o'l
Polymerization with t-DT Polymerization without t"DT [
10".(1] '" Initiator conversion in % after 10'.R"., conversion in % after lO'.R,ru 0 ....
mol.L' ; 2hr 4hr 6hr 8hr g.mL'.s·' 2hr 4hr {)hr 8hr g.ml:'.s·' 2.: '" PPS 13.0 15 39 63 80 1.5 3 6 7 7
ACPA 13.0 14 38 67 84 1.6 14 39 65 82 1.6 AlBN 13.0 14 37 62 80 L6 13 40 63 79 1.6
SPS 13.0 11 29 55 79 1.82 11 25 43 66 1.38 ACPA 13.0 14 32 60 82 1.77 SPS 0.52 19 45 69 83 1.64 13 29 46 61 0.99
ACPA 0.52 13 33 55 73 131
Tabie 8.4. Effect of the rype af emulsifier all reaction kinetics, both in the presence mui absence of t-DT."
[cation] Polymerization with t"DT Polymerization without t"DT
Emulsifier mol.L'
conversion in % after lO'.R,." conversion in % after 10'.R,.o, 2hr 4hr 6hr 8hr 24hr g.mL'.s·! 2hr 4hr 6hr 8hr 24hr g..rnL'.s·'
Dresinate 004 15 39 63 80 99 1.52 3 6 7 7 10 K-stearate 0.16 21 46 75 89 100 1.85 3 6 8 9 20 K"oleate 0.16 15 37 65 82 100 1.71 2 3 4 6 20 0.21b)
SDS 0.3 11 29 55 79 99 1.82 11 25 43 66 98 1.38 ... ... (Jl.
0) Initiator is persulfate; [I] ~ 13.10"' moLL"'. b) I\.. determined between 20 to 35 % conversion.
116
"if. 100
.f: c 0 80 5 € .~ 60 -.;;J
~ ~ 40 ::J
E ::J (j
20
a
"if. 100 .S c 0 80 . .::; ::J .0
~ "0 60
'" ;>
~ 40 ::J
E ;:, (j
20
0
Chal>ter 8
Table 8.5. Emulsifier-free emulsion polymerizations of butadiene.
a
C,' lO·~.C'_DT llf.~ 1W'.R.,oI/N d.. at 90 % L- t g.g-1 g.mL-' .s"1 g.S·1
P g. run
0.0 0.0 0.19 4.7 173 1.26 0.0 7.0 0.18 3.9 167 1.16 4.0") 7.0 0_79 4.9 173 1.03
&perimentaJ derails emulsifier-free rons.- M/W = 1/9 (w/w), [PPS] = 10 mmoI.L-1
, [~C031 -- 10 mmol.L-l, n .. 500 rpm . • ) Polymerization described in section 5.2.
25 ¥i-
," c 20 2
u
~ 15
10
5
0 50 100 150 200 250
d,/nm
orO' o-¢ 50 * / c:
" c I 402
U ., g / 30
Q Figure 8,]. The cumulative (0) and
20 fractional particle size distribution~
" versus number average diameter do for the emulsifier-free polymerization
/ 10 without t--d.tx/ecanethio{ (top;' and the ., poiymerizalion with 4,0 gL- SDS and ,i t-DT (bottom). For details see text.
0 100 150 200 250
d,/nm
Effects 0' 'J'hiols 117
The particle size and size distribution of the final latexes of tbese emulsifier
free polymerizations (Figure 8.3), reveal some otber marked peculiarities_
For the emulsifier-free emulsion polymerization of styrene Goodwin et at 12)
have derived an empirical correlation between the particle diameter, initiator
and monomer concentration, ionic strength, and temperature. From this relation
ship it follows that under the experimental conditions specified in Table 8.5, a
monodisperse polystyrene (PS) latex with a diameter of 700 nm would be
obtained. In contrast, the polybutadiene (JOB) latexes are highly polyuisperse,
and of a significantly smaller diameter. The large polydispersity strongly suggests
that (homogeneous) particle nucleation is not restricted to the very beginning of
the polymerization (i.e, interval l), and that once formed, the PB'particles are
highly stable toward coagulation. Consequently, the number of stabili:;:;ing 504-
-groups (introduced by the initiator) on PB-particles must be significantly larger
than on PS-particles prepared under identical conditions. This is consistent with
the observation that the rate of volume growth, being proportional to ~ and Ii,
for a PB-particle is significantly smaller than for an equally si:;:;ed PS-partide.
Although the socalled promoting effect is still not completely understood,
some aspects have become dearer, such as:
- The structure and purity of the emulsifier, as well as concentration and
type of initiator, all playa crucial role.
- There is no apparent reason why sulfate radical anions (or oligomers
thereof) should not initiate butadiene, as is shown in the persuLfate
initiated emulsifier-free polymerizations in the absence of thio!.
- Although this does not disprove the involvement of thiol radicals in the
initiation mechanism in polymeri:;:;ations with emulsifier, a simple redox
reaction between thiol and persulfate would oversimplify the actual situation.
The most pu:zzling aspect yet remains to be solved: why is the promoting
effect specific of diene-hydrocarbons, such as butadiene and isoprene, and not
showing up when using other monomers. Styrene can readily be polymerized
without a chain transfer agent, using any type of initiator or emulsifier (including
dresinate 214), both at widely varying concentrations.
118 Chapter 8
8.S Radical desorption
Regardless of the complexity of the data presented abo"l/e, one common
characteristic is the observation that addition of t-DT never decreases the steady
state 1\.01' nor the average rate per particle ~1/N. If a cbain transfer agent
were facilitating radical desorption, a decrease in Rpo/N is expected, whereas
all experimental data indicate the reverse.
Nomura et al. l ) investigated the effect of radical desorption by chain transfer
agents in the emulsion polymerization of styrene. Carbon tetrachloride and
tetrabromide, ethanethiol, and n-butanethiol were found to give extensive radical
desorption; in contrast, n-dodecanethiol (n-DT) did not affect the polymeri~ation
rate nor the particle number. The fact that n-DT radicals apparently do not
desorb was attributed to their extremely low water sOlubility. The concentration
of n-01' was varied between 2.7 lO-s and 1.1 10"" mol.g-1 styrene, comparable
with that in the prcsent system, viz. 3.5 10-5 mol t·DT per gram butadiene.
Table 8-6 Effect of various grades of dodecanethiol in polymerizatioNS with 32.4 g.L. 1 sodium dodecylsulfate (recipe 2).
thiol conversion in % after
2hr 4hr 6hr
t_DT'l 11 29 55 t-Dr) 11 29 56 n·OT 13 32 62
oj t-DT supplied by Pennwalt. 0) t-DT supplied by Fluka.
8hr
79 79 83
ItY.Rpol d cJ 90 x'
g.mCls·! nm
1.82 46 0.598 ),82 44 0.613 1.88 S1 0.608
c) ~ at 90 % conversion.
A final check was made in relation to the quality of the commercial grade
t-DT (supplied hy Pennwalt), a crude mixture of en-isomers which may contain
small amounts of lower and higher thiok Additional polymerizations were
performed (Tahle 8.6), using a mixture of only two teniary C12-isomers (supplied
by Flub; for structures see section 3.6) and pure n·DT. Both polymerizations
with t-DT were found identical, while the one with n-DT Willi only slightly faster.
ElTects of Thiols U9
This latter observation is in aJ:J:ord with results reported by Kolthoff.3) However,
allowing for the difference in particle size and the dependence of ii on particle
size, the average rates per particle are equivalent for all three Cl2"thiols used.
Radical desorption of dodecanethiol radicals is obviously negligible, regardless
of the structure and composition of the dodecanethiol employed.
8.6 Conclusions
From a kinetic standpoint, t-dodecanethiol appears to behave as an 'ideal'
chain transfer agent in the emulsion polymerization of butadiene; it regulates
cross-linking efficiently, but otherwise does not affect the reaJ:tion kinetics:
" Eventhough the gel point varies widely with the initial amount of thiol,
CM and )C' were both found constant within experimental error. Even in the
absence of a chain transfer agent, the extent of cross-linking is still to low
to restrict the monomer swelling of latex particles.
- Dodeca.nethiol radicals do not desorb, regardless of their structure and
composition, as a result of an extremely low water solubility.
The promoting effect of slightly water-soluble thiols appears specific of diene
hydrocarbon monomers, but the phenomenon is stilI poorly underStood.
Nevertheless, some aspect can now be placed in a better perspective:
- Sulfate radical-anions (or oligomers thereof) are capable of initiating the
emulsion polymerization of butadiene, as was shown in the emulsifier-free
experiments. which show polymerization with an Rl"'iN identical to that of
reactions with emulsifier and thiol leading to the same final partide size.
- Despite tbis observation, many authors (including the present one) have
reported On the retardation of butadiene emulsion polymerizations in the
absence of an appropriate thiol, using different types of emulsifiers.
- The extent of retardation in the absence of thioi depends on the structure,
concentration and chemical purity of the emulsifier and initiator employed.
- It has become clear that a simple redox reaction between a sulfate radical
anion and ihiol is inadequate to provide a satisfactory explanation.
120 Chapter 8
Appendix Anomalous effects with dresinate 214 as emulsifier
A strange phenomenon was encountered woen particle numbers were
compared of polymerizations with dresinate 214 in the presence (initiator PPS)
and absence (initiator ACPA) of t-DT (Figure 8.4, extension of Figure 4.5). On
the basis of the previous discussion the choice of the initiator at [I] = 13.10--'
moLL·1 and the absence of t·OT is not expected to cause any disparity in the
final particle number. This is indeed found for all polymerizations with C,\ .:: 32.4
g.L·1• However, the polymerization with C" = 64.8 g.L.1 is markedly different
with regard to the final N; without t-DT the particle number is significantly
larger (N = 9.1015 mL·1) than in its presence (N '" 3.1015 mL·1; see Figure 8.4».
The same phenomenon was encountered again (although less pronounced), when
lowering the monomer/water weight ratio in recipe 1 from 3/7 to 2/8, keeping
all othl;f ingredients constant on water basis.
~ 3.5
E ---... Z 3.0 ~
• ,,-_---=."'-- .. -- '" _ .. --", b
2.5
2.0
1.5
1.0
0.5
00 0 20 40 60 80 100
Conversion in %
Figure 8.4. Development of pMicie number N with conversion Jor polymerizations in tile presence (open symbols) and ahsence of t-DT
(full Jymb(}L~) at different dresinale 214 concentrations Crc -- 64.8 g.L"i (6); C. i= 32.4 gLI (0); CF. '" 16.2 gLI til); C£ -- 8.1 gL,I (0).
ElTects of 'lbiols 121
This effect was studied more systematically by varying the amount of t-DT
using ACPA as initiator ([I] .. 13.10-3 mol.L-I ), and determining the final particle
diameter at 90 % conversion (Figure 8.5). A steady increase in particle diameter
~ is observed. Figure 8.5 also contains data points for both standard recipes
with dresinate 214 and sodium dodecylsulfate. For the polymerizations with 64.8
gL-1 dresinate 214, the conversion-time curves (Figure 8.6) and the particle size
distributions (PSD) of the final latexes (Table 8.7) were also determined.
E .:: "'-
:;! 100 ~.:.~,,/o Ll
80 Figure 8.5. Panicle diameter dw
o t:I at 90 % converyion (dr;o) vs.
/ /
concentration ofilliol (C'_Dr)' for polymerizations with C ... 64.8
o gLI dresinate 214 (0); c~ '" 60
r~ 32.4g.L-I dresinate 214 (4); CF.
= 32.4 g.L-1 SDS ('17); (1J'Y} C ... 32.4 gLJ tiresinate 214 (01 Wilh
_~_~--1jI va monomer/water ratio" 2/B. 40-+~ ........... ---.---.---r---r-...,...---' (ACPA] = 13.1rr molL-I.
0.0 0.4 O.S 1.2
Table 8. 7. Data on panicle size distributions oj latexes prep(1Ted wilh 64.8 g.L·1 dresiltale 214 and varying amounts of t-dodectlnethiol t-DT.
16'.Ct_DT C in % do d,. dmin d,.o. sd P g,g'! nm run nm nm nm
0.0 99.7 45 50 25 69 8 1.1 0.7 99.0 63 76 21 104 20 1.2 1.4") 99.2 88 112 32 140 29 1.3 2.8 99.4 64 103 25 170 30 1.6
Abbreviations: dmio and dmax are the minimum and maximum particle diameter do, respectively, and sd the standard deviation. a) Particle size distribution slightly bimodaL
122
<fi 100
,5 c a 80 '~ ill > C 0 60 0
40
20
0
0 2 4 6 8 Time in hours
Chapter 8
Figure 8.6 Conversion v.~. time curves Jot polymerizations wirh C. = 64.8 g.L-J dresinate 214 and II] '" no mmo/.c l ACPA, with varying amounts oj /-DT CI-DT " 0.0 g.g.1 (6); C'_DY .. 1.4 10-1 g.,/ (0); and C'-DT = 2.8 J(Jz g.g-I (\1).
Increasing the amount of t-DT clearly prolongs the particle nucleation
period, resulting in a broadening of the PSD_ Obviously, we are deaJing with
an exceptional behaviour, since this phenomenon was only encountered at a
relatively high dresinate 214 to butadiene ratio, and not at all with SDS at
comparable particil; sizcs_ Radical desorption or cross-linking cannot provide a
plausible explanation, as is evident from the dis\:ussion in the preceding sections.
It may very well be that, as in the Case of the promoting effect, impurities
present in dresinate 214, interact with t-DT and/or pen;ulfate, thereby
influendng the radical entry rate and thus the particle nucleation rate in these
ab illirio experiments. Further research is necessary to unravel the exact rok of
thiols of low water solubility in the mechanism of radical entry, especially in
emulsion polymerizations of diene-hydrocarbons.
References 123
M. Nomura, Y. Minamino, K Fujita, M. Harada 1. Polym. Sci., Polym. Chern. Ed. 20, 1261 (1982)
2 G. Lichti, D.F. Sangster, B.C.Y. Whang, D.H. Napper, R.G. Gilbert, 1. Chem, Soc., Faraday TraJIS. ] 78, 2129 (1982)
3 I.M, Kolthoff, WE Harris,l. Polym. Sci. 2, 41 (1947) 4 P.A Weerts, J.L.M. van der 1.oos, AL German, Polym. Comrn, 29, 278
(1988) 5 M. Morton, S. Kaizerman, M.W. Altier,1. Colloid Sci- 9,300 (1954) 6 PJ. Flory, J. Rehner, 1. Chern. Phys. 11, 521 (1943) 7 F,A. Bovey, I.M. Kolthoff, A.!. Mec;lalia, E.J. Meehan, "Emulsion
Polymerization", Interscience Publishers, New York 1955, p. 344 8 M. Morton, JA Cala, I. Piirrna, J, Polym, Sci- 15, 167 (1955) 9 I.M. Kolthoff, I.K Miller,J. Am. Chern. Sor::. 73,5118 (1951)
10 K. Ishigure, T. O'Neill, E.P. Stahel, V. Stannert, 1. Macromol. Sci. - Chern. A8(2), 353 (1974)
11 R.s. Bhakuni, PhD. Thesis, University of Akron (1964), Chern. Abstr. 62, 13351b (1965)
12 J.W. Goodwin, J. Hearn, c.c. Ho, R.H. Ottewill, Colloid Polym. ScL 252, 464 (1974)
Summary 125
SummaJ')'
Emulsion polymerization is an extremely complex process, as it comprises
a heterogeneous reaction system, consisting of three distinct phases: the aqueous
phase, a colloidal dispersion of polymer particles, and a monomer emulsion.
Radical polymeri~ation may commence simultaneously in any of these phases,
while exchange of material between the phases always needs to be considered.
This thesis attempts to provide a consistent overview of the effects of all
relevant chemical, colloidal and physical parameters on the emulsion polymeriza
tion of butadiene. The reason for selecting butadiene is twofold: (1) although it
was one of the first monomers to be studied systematically, very little is known
about its kinetic behaviour in the context of recent models and theories; (2)
polybutadiene containing polymers are of significant commercial importance.
As a starting point was chosen a typical industrial recipe, including the
commercial emulsifier dresinate 214 (chapter 4). A study of the effects of
initiator ([I]) and emulsifier concentration ([E]) led to the following preliminary
conclusions: (1) initiator efficiency appears to be low; (2) particle fonnation is
coagulative in nature; (3) the average number of radicals per particle Ii 1- 0.5.
These observations were further tested by performing analogous reactions
with sodium dodecylsulfate as emulsifier (chapter 5). It was shown that the
coagulation was caused by the high ionic strength of the aqueous phase. K:inetic
relationships between particle number, and [I] and [E], take a complex form
when coagulation is involved. The dependence of average rate per particle
(equivalent to Ii) on particle size was found identical for both emulsifiers used,
despite the enonnous differences in composition, purity and performance.
126 Summary
Being a heterogenous system, emulsion polymerization is susceptible to the
agitation conditions (chapter 6). The following effects were established: (1) at
high ~tirring speed~ (equivalent to impeller Reynolds number Re > Iff) the
effective emulsifier concentration, and consequently the final particle number,
decrease significantly due to emulsifier adsorption onto the monomer droplets;
(2) at low stirring speeds the polymerization becomes diffusion-contrOlled, a.~ a
result of a (partial) phase separation of the monomer emulsion. Simultaneously
the particle number increases as less emulsifier i~ ad~orbed Onto the droplets.
At monomer volume fractions > 0.6, additional effects occur as mixed monomer
emulsions are formed, affecting both particle number and polymerization rate.
'''he combined kinetic information of intervals II and III of the ab initio
experiments discussed above, unambiguously showed (chapter 7) that n < 0.5
(except perhaps for very large particles). This implies that bimolecular
termination within the particles is not rate-determining, and that radical exit
from the particles is significant. The latter is most likely to be facilitated by
oligomeric butadiene radicals. The low ii behaviour, in combination with a low
initiator efficiency, explains the small effect [lJ has on polymerization rate.
La.~tly, the role of Cn-thiols was considered. Apart from being effective
chain transfer <lgent~, no additional effects were found: (1) monomer solubility
within the particles was found independent of the amount of thiol, and (2) the
corresponding tbiol f<ldicals do not exit due to their low water solubility. With
regard to the rate-enhancing promoting effect, it was found that impurities
present in other recipe ingredients strongly determine its magnitude. In
persulfate-initiated emulsifier-free polymerizations the promoting effect was
found to be completely absent.
Based on this work, further research may be focused on the following items:
(1) a full kinetic investigation of the emulsion polymerization of butadiene,
preferably using monodisperse seed latexes, in order to obtain reliable values for
the rale coefficients of propagation, exit, and possibly termination; (2) a
mechani,tic interpretation of the first-order radical loss process; (3) to provide
an adequate e'lplan<ltion for the puzzling promoting effect of thiols in diene
hydrocarbon polymerizations,
Samenvatting 127
Samenvatting
Emulsiepolymerisatie is een uitermate complex proces, aaar het een
heterogeen reactiesysteem vertegenwoordigt, bestaande uit drie verschillende
fasen: de waterige fase, een colloidale dispersie van polymeerdeeltjes en een
monomeeremulsie. Radicaalpolymerisatie kan gelijktijdig in elk van deze fasen
plaatsvinden, e'lenals een continue uitwisseling van materiaal tussen de fasen.
Oit proefschrift beoogt een min of meer compleet over:dcht te geven van de
invloed van alle relevante chemische, colloid ale en tysische parameters op de
emulsiepolymerisatie van butadieen_ Hierbij is VOOT butaaieen gekozen, omdat:
(1) ondanks het feit dat butadieen een van de eerste monomeren is dat systema
tisch is bestudeerd, is van het kinetische gedrag nog steeds weinig hekend: (2)
polybutadieen bevattende polymeren van grOOI cornmercieel belang zijn_
Als uitgangspunt werd een typisch industriele receptuut gekozen, met de
com.merciele emulgator dresinaat 214 (hoofdstuk 4). Vit de invloed van de
initiator- ([I]) en emulgatorconcentratie ([E]) konden de volgende voorlopige
conclusies worden getrokken: (1) de initiator efficiency is laag; (2) deeltjes
vorming is coagulatief; (3) het gemiddeld aantal radicalen per deeltje Ii f 0,5.
Deze waarnemingen werden geverifieerd door analoge experimenten uit te
voeren met natrium dodecylsulfaat als emulgator (hoofdstuk 5). Hierbij is
aangetoond dat de coagulatie veroorzaakt wordt door de hoge ionensterkte.
Kinetische relaties tUSSen het aantaI deeltjes, [I] en [E], nemen hierdoor een
compleILe vorm aan. De afbankelijkbeid V".I.n de gemiddelde snelheid per deeltje
(equivalent met ii) van de deeltjesgrootte was identiek voor beide emulgatoren,
ondanks hun gtote vetschil in samenstelling, zuiverheid en performance.
U8 SaOlenvlllting
Vanwege het helerogene karakter zijo emulsiepolymerisaties gevoelig voor
de roercondities (hoofdstuk 6). De volgem.le effecten zijn waargenomen: (1) bij
hoge roersnelheden (met een Reynoldsgetal Re > 1(Yl) nemeo de effectieve
emulgatorconccntratie en het aantal deeltjes significant af, tengevolge van
adsorptie van emulgator op de monomeetdruppels; (2) bij lage roersnelheden
is de polymerisatie diffusiegelimiteerd, vanwege een (gedeeltelijke) fasen
scheiding van de monomeeremulsie. Bij monomeer volumefracties > 0.6 treden
additionele effecten op, doordat gemengde monomeeremulsies worden gevormd,
hetgeen zowel het aantal deeltjes als de polymerisatiesnelheid bcinvloedt.
Pe gecombineerde kinetische informatie uit de intervaHen 1I en HI van de
hjerboven beschreven ab initio experimenten (hoofdstuk 7) toont onomstntelijk
aan dat n <:: 0,5 (behalve misschien voor grote deeltjes), Dit betekent dat
bimoleculaire terminatie in de latexdeeItjes niet snelheidsbepalem.l is, tetwijl
radicaaldesorplie aanzienlijk moet zijn. Het laatste gebeurt vermocdelijk door
oligomere butadieenradicalen. He! lage n-gedrag, io combinatie met een lage
initiator effkiency, verklaart het geringe effect van [I] op de reactiesnelheid.
Als laatste is de rol van C1z-thiolen bekeken. NailSt de functie van effectief
chain transfer agent. zijn hierbij geen additionele effecten waargenomen; (1) de
monomeeroplo~baarhcid in de latexdeeJtje~ is onafhankelijk van de hoevee1heid
thiol, en (2) thiol radicalen desorberen niet vanwege hun geringe water
oplosbaarheid. Met bctrekking tot het snelheidsverhogende 'promoting effect'
van lhiolen is gevoo(len dat dit medebepaald wordt door verontreinigingen
aanwezig in de andere ingredienten. In persulfaat-geinitiecrde emulgatorvrije
polymerisaties bleck het promoting effect volledig afwezig.
Geb<l$eerd op dit werk, wu toekomstig onderwek zich op de volgeode
gebleden kunnen richten: (1) een volledig kinetisch onderzoek van de emulsie
polymerisatle van butadicen, bij voorkeur met monodisperse seedlatexen, gericht
op het bepalen van nauwkcurige waarden vOOr de snelhcidscoostanten van
propagatic, desorptie en terminatie; (2) cen mechanistische intcrpretatie van het
ecrste-orde radicaal verUes proces; (3) een verklaring van het promoting effect
van thinlen bij de emulsiepolymeri~atie van dieen-koolwaterstoffen.
Glossary of Symbols 129
Glossary of Symbols
Symbol 51-units
a dimensionless parameter, defined by a '" (&f!j
A surface of a monomer-swollen particle m2
A.:- total interfacial area per unit volume of dispersed phase m-I
A.J specific area per emulsifier molecule m2
c (fractional) conversion
c" emulsifier concentration on weight basis kg.m-3
~ monomer concentration within a particle mol.m-l
C,.Q initial monomer concentration within a particle mol.m·3
C.q monomer concentration in the aqueous phase mol.mo3
d average particle diameter m
d .. average monomer droplet diameter m
d"" swollen particle diameter m
d .. diameter at xx % conversion m
D impeller diameter m
Dr reactor diameter m
Dp diffusion coefficient in a pattide m2.s-1
0 ... diffusion coefficient in the aqueous phase m2.S-1
e fundamental electronic charge A.s
g gravitational constant m_s·2
H., volume swelling ratio
I Bessel function of the first kind
k (first-order) rate coefficient for radical ellit S·I
130 Glossary of Symbols
r,He coefficient for radical ahwrption
rate coeWcient for initiator decomposition
propagation rate coefficient
nlte coefficient for bimolecular termination
rate coefficient for transfer to monomer
termination rate coefficient in the aqueous phase
Boltzmann constant
m dimen~i()nless parameter, defined by m = k.v/k,
M, average molecular weight of segments between
two cross-link~
n
number-average molecular weight
weight-average molecular weight
monomer to water weight ratio
stirring speed
n average number of radicals per particle
n; number of particles with diameter d;
n 0 M
N
amount of monomer present at the beginning of
interval III, per unit volume of aqueous phase
amount of monomer initially present per unit
volume of aqueous phase
particle number per unit volume of aqueous phase
Nn number of particles containing n free radicals
Avogadro's number
polydispersity index, defined by P = ~/dn q partition coefficient of exiting species hetween
particles and the aqueous phase
R rate of radical production (per unit amount of
emulsion), assuming 100 % initiator efficien(''Y
R gas constant
Rp<'] rate of frce radical polymerization
Rp',JN average rate per particle
Re Reynolds impeller number, defined by Re
S·I
5.1
m3.morl.s'1
m\,mor1),s·1
m 3.mol"_s']
m3.morl.s·1
kg,m2.K'I.:;'~
kg,m2.K·1_mor1.S·2
kg(or mOI)_m--3.s·1
kg(or mol)SI
Glossary of Symbols 131
settling time of a dispersion
to reference settling time
tcr critical time, duration of interval I
t l / 2 half life time s
t,.. period of time required to reach xx % conversion ~
T absolute temperature K
v volume of a (monomer-swollen) particle m]
V total volume of polymer per unit volume
of aqueous phase
V., monomer molar volume m3.morl
wp weight fraction of polymer (on a monomer-
polymer basis)
x fractional conversion in interval III
x' (fractional) conversion where interval III begins
y dimensionless parameter, defined by Y = 2Nk1k1./k. ~
z parameter in the Nomura transfer/diffusion model
cl< dimensionless parameter, defined by cl< ;;; p.v/Nkt
cl<' dimensionless parameter, defined by cl<' '" PjvjNk,
i interfacial tension kg.s-2
eo pennittivity of free space kg.m.K2.s·~
e. relative permittivity
~ Zeta potential kg.m2.A-1.s·3
temperature in degrees Celsius K
{J particle volume growth rate in the Smith-Ewart theory m]_s·t
/I' parameter in the Debye-Hucke\ theory m·1
/-I (average) dynamic viscosity kg.m·\s·1
/I kinematic viscosity, defined by /I '" "jp m2.s·!
p (pseudo-first-order) rate coefficient for radical entry S-I
p. (average) density of substance Or phase x kg.m·l
/l.p density difference kg.mo3
p. rate of radical entry m-'.s-l
132 Glossary of Symbols
Pi rate of radical production in the aqueous phase m·].s·l
a ionic strength moLm-3
'P volume fraction
+ .. monomer volume fraction within a particle
X Flory-Huggins interaction parameter
"'0 surface potential of a particle kg.ml .A1.s-3
"', potential at a distance r from the centre of a particle kg.m2.A-l .s-3
[El emulsifier concentration moLm-:)
[EJ<rr effective emulsifier concentration moi.m-3
[I] initiator concentration mol.m-3
[MJ monomer concentration moLm-3
[R·J free radical concentration mol.m~'
['11 intrinsic viscOSity m3.kg-1
Subscripts and :.bb~iations:
aq aqueous phase n number-average
c continuous phase p polymer, propagation
D dispersed phase surface-average
B emulsifier v volume-average
I- latex w weight-average
M monomer
Curriculum Vitae 133
Curriculum Vitae
De schrijver van dit proefschrift werd op 23 augustus 1959 in Breda geboren"
Hij behaalde in 1976 het havo diploma en in 1978 het atheneum-B diploma,
beide aan de Newman-Ypelaar schoiengemeenschap in Breda.
In 1978 began hij aan de studie Scheikundige Technologie aan de
Technische Universiteit Eindhoven, alwaar het kandidaatsexamen met lof werd
afgelegd in februari 1982. Het doctoraalexamen met lof, tesamen met de
onderwijsbevoegdheid voor 5cheikunde, werden behaald in ~eptember 1985, na
een afstudeerwerk op het gebied van de synthetisch organische chemie onder
leiding van prof. dr. E.F. Godefroi.
In oktober 1985 werd een aanvang gemaakl met het promotieonderzoek
naar de emuIsiepolymerisatie van butadieen in de vakgroep Polymeen;hemie en
Kunststoftechnologie (TUE), onder leiding van prof. dr. it. AL German. In het
kader hiervan vcrbleef hij drie maanden aan de School of Chemistry, Sydney
University, in de groep van prof. dr. RO. Gilbert en prof. dr. D.H. Napper.
Per 1 januari 1990 trad hij in dienst bij DSM Research in Geleen, als
researchmedewerker op de afdeling CP-harsen.
134 Dankwoord
Dankwoord
Bij deze wil ik iedereen bedanken voor de getoonde belangstelling, de
wetenschappelijke bijdragen, de ondersteuning op techniscb en administratief
gebied en nie! in de laatste plaats voor de warme collegiaIiteit, waaTdoor ik met
veel pie:l;ie. en enthousiasme aan mijn promotie heb \mnnen werken.
Stellingen
(behorende bij hef proefschrift van P.A. Weens)
Het bepaJen van de oplosbaarheid van butadieen in water bij drukkeu kleiner
dan 1 atmosfeer gaat volledig voorbij aan het feit dat met dit gasvormige
monomeer altijd onder verhoogde druk wordt gewerkt. J.w. Melloin, JJ. o·Cu."",.l Am. Chim. StJe. 6.1. S7S (1941)
Eenmaal gepubliceerd kunnen data hardnekkig blijven circuleren, Ongeacht hun
betrouwbaarheid of relevantie. Dit blijkt ondermeer ult recente literatuur
waarden voor de wateroplosbaarheid van butadieen, welke allen teruggaan op
het werk van McBain en O'Connor, echter zonder specificatie van de originele
cxperimentele condities. F.A. Bovey. I.M. Kol1boff. A.I. M.;i;I.IIU.\I, E-J. M'~~b.;J:n. in "'~mumDR Polymmuman", 1I11C.~ic.nu, New Vork 19l5, p.1M
W.M. SaUman., in "EnC)'Ctopedia 0/ Polymer Scierw.e.QN/ T~~", rl1t1;~ien~, New York 19M, ht ed., vol. 2. p. 682 WJ. Bail.ey. il3! 'VInyl and DkMM,,~H. ~ir~ 'by E.C L¢ortlllN. Wilq_In~r.;ci~n~. N~ York 1971, p. 798
D.C. B1ackJey. "Emubion Po/y.....wrlo. - 'l'1owry <RId p,,,,,w..". Appliod P.bli.lI< ... Loodon 197~, p.44! O.W. f'oobl<ln, I. "Enu<IaI<m ~', oditod by I. Pii",... /I.ond.m., P=<, N .... Vork 1982. p.372 J.w. Va.dethon. J. PoIym. Sci. PoIym. !lJ!mp. 72. 161 (\985)
Het gebruik van de zogenaamde 'bottle polymerizer' voor kinetisch onderzoek
aan heterogene systemen is ten sterkste af te raden, dit in verband met de zeer
ineffidente menging. Hoor~t\1k: IS ~i~ dit pmefg:hrift.
Inconsequent gebruik van de eenheden gram en mol in een en hetzelfde artikd
kan leiden tot grote verwarring en komt derhalve de wetenschappelijke betrouw
baarheid niet ten goede. M. Nom .... Y. Mi""min<:>, ... fujita. M. ".nuIa, I. Po/ym. Sci., 1'oI)m. Chom. Ed. 20. 1261 (1982)
De beschrijving van de eopolymerisatiekinetiek van styreen en acrylonitril in
aanwezigheid van polybutadieen door Hagiopol et al., gaat volledig voorbij aan
het heterogene karakter van de gevormde ABS-resin. c. Haglopol, T. Del ... o. T. Momoteo. I. AppI. PoIym. Sci. )7, 941 (1939)
De modellering van de emulsiepolymerisatie van butadieen zoals beschreven
door Morbidelli et al. laat overtuigend :1;ien dat een goede data-fit nog niet
betekent dat het gebruikte model ook correct is. M. Morbldelli. G. Siorti. s. Cam. J. AppI. Polyrn. Sci. U. 901 (1983)
"Onhekend maakt onbemind". De groo!schalige industd~le toepassing en lang
durige historische onLwikkeling van butadieen in emulsiepolymerisatieprocessen,
suggereert, geheel ten onrechte, een grote bekendheid met de kinetiek van dil
monomeer. Hooldslukk •• I 1/«> S "0' dh proor",htifi.
1'erugdringen van hct au!ogchruik middels financiele maatregelen zal weinig
effect sorteren, omdat de automobilist be reid is een (te) hoge prijs voor zijn
individuele vtijheid te betaltn.
Herste1 van de Oostduitse economie Iijkt het meeste gebaat bi] teruggave van
alle onteigende goederen aan de voormalige eigenaren (vooropgesteld dat ze die
rotzooi uberhaupt nog terug willen).
Afschaffing van de wettelijke bescherming van aile univeTsitaire titeh met
uitzondering van de doctorstitel, zoals voorge~teld door de minister van
onderwijs, is wellicht een probaat middel om het pWmoveren wat aantrekkeIijker
Ie maken. UiI.pmok mini.,., Ril .. n, NRC Tf~",,<4N~(i ~.~. 13 april 1990.
Tijdens congressen en symposia overtreft het aantal doctoren altijd het aantal
gepromoveerden.